Set the chart's placeholder version to 0.37.0

replicas and strategy: Recreate were the chart's only guard against the
archiver, notifier and migration races; v0.36.0 closed all three with
advisory locks and a conflict-resolving incident insert, which made
that guard redundant rather than load-bearing. Expose replicaCount
(new, no values.yaml key existed before) defaulting to 2, and switch
to strategy: RollingUpdate with no explicit maxUnavailable/maxSurge --
the 25%/25% default rounds to 0/1 at replicaCount: 2, which is already
zero-downtime.

The chart does not gate this on image.tag, so pointing it at a
pre-v0.36.0 image with the new default is a foot-gun by omission --
noted in both the values.yaml comment and the deployment.yaml comment,
not guarded in code, same as the chart does for every other
version-coupled assumption today.

Co-authored-by: Claude <noreply@anthropic.com>

README.md

@@ -793,7 +793,7 @@ attempted.
 | `POST` | `/api/oidc/device/token` | `{"device_code"}` → `202 {"status":"pending"}`, then `200` with the session cookie once approved (once only). `410` with `{"error":"expired"}` or `{"error":"denied"}`; `429 {"error":"slow_down"}` if polled faster than `interval` |
 | `POST` | `/api/oidc/device/approve` | **session** — `{"user_code"}`. Approves a pending device login as the caller. `403` for an API key; `404` for an unknown, expired or already decided code |
 | `POST` | `/api/oidc/device/deny` | **session** — `{"user_code"}`. Refuses it |
-| `GET` | `/api/oidc/callback` | Where the provider sends the browser back. Sets the session cookie and redirects to `/`, or to `/?sso_error=<code>` — one of `denied`, `expired`, `failed`, `unavailable`, `not_allowed`, `no_email`, `email_conflict`, `disabled` |
+| `GET` | `/api/oidc/callback` | Where the provider sends the browser back. Sets the session cookie and redirects to `/`, or to `/?sso_error=<code>` — one of `denied`, `expired`, `failed`, `unavailable`, `not_allowed`, `no_email`, `email_conflict`, `disabled`, `not_bootstrapped` (no user exists on this install yet — sign in again once something has called `/api/bootstrap`) |
 | `POST` | `/api/logout` | Ends the session and clears the cookie |
 | `GET` | `/api/me` | The caller: `{user, has_password}` |
 

SERVICE-ACCOUNTS.md

@@ -119,38 +119,100 @@ new key, revoke the old one," not "recreate the account."
 
 ### Auth middleware
 
-`internal/api/middleware.go`'s existing dual resolution (`Authorization: Bearer`
-→ `apiKeyUser()`, or session cookie → `sessionUser()`, both landing on the same
-`models.User` + team-membership context) gains a third path: a bearer token that
-hashes to a `service_account_keys.key_hash` resolves to a distinct principal
-type, not a synthesized `models.User`. `requireTeamMember`/`requireTeamOwner`
-treat a matching team-scoped service account as owner-equivalent for that one
-team (satisfies the same checks a real team owner would), and an instance-scoped
-one as satisfying `AdminOnly` for team-creation/listing purposes **and** for
-minting a `team`-scoped service account against any team (`POST
-/api/service-accounts {"scope":"team","teamID":...}`) — this second permission
-is what lets an operator-style caller create a team, then immediately mint that
-team its own narrower credential, without a human in the loop for every team.
-Neither permission extends to user-management endpoints (`POST /api/users`,
-`PUT /api/users/{id}/admin`, etc.), which stay human-admin-only. Anywhere
-identity is recorded for a human (incident timeline
-`acknowledged_by`/`assigned_to`, audit-relevant fields), a service-account
-principal is stored and displayed distinctly, e.g. `service-account:terdut-operator`,
-never coerced into a `user_id` FK.
+**Revised** (this section originally described an aspiration that didn't
+match what shipped — `TEAM-LOOKUP.md` already caught one instance of that,
+and a fuller audit found three more; this is the corrected, as-built
+description, not the original proposal).
 
-**Team scope, as implemented, is owner-equivalent for every `requireTeamOwner`
-endpoint, membership and invites included — nothing server-side carves those
-two out.** That's broader than what `terdut-operator`'s CRDs actually need
-(escalation/deadman/integrations/OIDC-bindings only; membership is explicitly
-never gitops-managed, see its DESIGN.md §4.2), a gap acknowledged rather than
-closed here: narrowing this to exclude
-`POST/DELETE /api/teams/{teamID}/members*` and
-`.../invites*` specifically for a service-account caller is a small, isolated
-follow-up (special-case those handlers rather than `requireTeamOwner` itself,
-which every other owner-gated endpoint still wants shared). Until then, what
-actually keeps membership out of automation's hands is that no operator built
-against this scope should ever call those two endpoints — not a server-side
-refusal.
+`internal/api/middleware.go`'s dual resolution (`Authorization: Bearer` →
+`apiKeyUser()`, or session cookie → `sessionUser()`) and the service-account
+path (`serviceAccountFor()`) both resolve into one `Caller` type
+(`internal/api/caller.go`), not two parallel, un-unified context
+representations the way an earlier version of this server kept them. Every
+authorization predicate reads `Caller`'s methods:
+
+- `Caller.IsAdmin()` — true **only** for a human system administrator, never
+  for a service account of either scope, under any circumstance. `AdminOnly`
+  and `requireSelfOrAdmin` key on this alone — user management
+  (`POST /api/users`, `PUT /api/users/{id}/admin`, etc.) and
+  `GET/PUT /api/admin/settings` stay human-only, forever. The original text
+  here claimed an instance-scoped service account satisfies `AdminOnly` "for
+  team-creation/listing purposes" — that was never true of the shipped code
+  (`TEAM-LOOKUP.md` caught the listing half; the creation half was always a
+  separate, bespoke check in `handleCreateTeam`, not `AdminOnly` itself) and
+  is not being made true now. Don't widen `AdminOnly`: every time this has
+  come up, the fix has been a narrower, purpose-built capability instead
+  (`?name=` lookups for teams and service accounts; now
+  `terdut-operator`'s own invite-minting feature for the one real gap this
+  boundary left — how a human ever gets a first login on a no-OIDC,
+  operator-managed install. See the bottom of "What this unblocks.")
+- `Caller.IsInstanceServiceAccount()` — true only for an instance-scoped
+  service account, never for a human (including a human admin).
+  `handleCreateTeam` uses exactly this: a human creates a team by being a
+  human (and becomes its owner); an instance-scoped service account creates
+  one with no human owner at all. The two paths are not interchangeable, so
+  this predicate deliberately does not also admit a human admin.
+- `Caller.Role(teamID)`/`TeamIDs()` — a human's real `team_members` rows, or
+  a team-scoped service account's single synthetic owner membership
+  (`serveAsServiceAccount`). This is what makes `requireTeamMember`/
+  `requireTeamOwner` treat a team-scoped service account as owner-equivalent
+  for that one team, with no separate branch needed in either function.
+- `Caller.ServiceAccountID()` — used by `OperatorModeBlock` ("any service
+  account passes") and by `callerMayManageServiceAccount`'s self-rotation
+  check.
+- `Caller.AsHuman()` — the accessor every handler that needs a real
+  `user_id` to act on behalf of must call and check, instead of reading a
+  user off context unconditionally. Before the `Caller` type existed, four
+  handlers did the latter and silently misbehaved for a service-account
+  caller: `handleMe` and `handleTestNotification` 500'd (a zero-value user id
+  that matches no row), `handleDismissOnboarding` silently no-op'd (`UPDATE
+  ... WHERE id = 0` affects nothing, still returns 204), and
+  `handleCreateInvite` wrote that same zero value into `invites.created_by`
+  — a real foreign-key violation, not just a wrong answer, since that column
+  is nullable but was never passed as `nil`. All four now call `AsHuman()`
+  and return an explicit 403 ("this endpoint is for human accounts only")
+  or, for the invite case, leave `created_by` `NULL` the same way
+  `handleCreateServiceAccount` already did for the analogous situation.
+
+**Team scope is owner-equivalent for every `requireTeamOwner` endpoint,
+membership and invites included — by design, not by an unclosed gap.** An
+earlier version of this document flagged this as "acknowledged rather than
+closed," kept in check only by the social convention that nobody *builds*
+automation against those two routes. That convention is retired:
+`terdut-operator`'s `TerdutTeam` controller now mints and revokes its own
+team's invite link through exactly this capability (its existing
+team-scoped credential, `POST`/`DELETE /api/teams/{teamID}/invites`), which
+is the real fix for the human-onboarding gap below — not a narrower
+carve-out of this capability. `service_accounts_test.go`'s
+`TestServiceAccount_TeamScopeManagesItsOwnInvites` pins it.
+
+**A team-scoped account can also mint another service account scoped to its
+own team** (`handleCreateServiceAccount`'s `callerOwnsTeam` branch, which a
+team-scoped caller already satisfies for its own team via the synthetic
+membership above). Kept, not restricted, for the same reason: a team-scoped
+credential is that team's owner's reach, full stop — carving this one
+capability out while leaving membership/invites alone would be an arbitrary
+asymmetry. Pinned by
+`TestServiceAccount_TeamScopeCanMintAnotherAccountForItsOwnTeam`.
+
+**`callerMayManageServiceAccount` gained the one load-bearing fix this
+redesign exists for:** an instance-scoped service account may manage
+(mint/revoke a key on) *any* team-scoped account, not only one admin, that
+team's human owner, or the account itself. `handleCreateServiceAccount`
+already let an instance-scoped caller *create* a team-scoped account for
+any team; this closes the gap where adopting or rotating one it didn't just
+create in the same call — exactly `terdut-operator`'s documented
+adopt-on-409 crash-window recovery (its own `DESIGN.md` §5) — 403'd forever
+instead of succeeding (`terdut-operator#3`). Pinned by
+`TestServiceAccount_InstanceScopeAdoptsAnExistingTeamScopedAccountsKey`.
+
+Anywhere identity is recorded for a human (incident timeline
+`acknowledged_by`/`assigned_to`, audit-relevant fields), a service-account
+caller is still coerced into a bare `user_id` of `0` today — `Caller`'s new
+`Identity()` accessor exists for exactly this follow-up, but wiring it in
+needs a schema migration (an actor-attribution column distinct from
+`user_id`) and is deliberately out of scope here. Tracked separately, not by
+this document.
 
 ## What this unblocks
 
@@ -177,6 +239,21 @@ Directly resolves `terdut-operator` DESIGN.md §6's two broken assumptions:
    holding many credentials in one place (the operator's namespace) an
    acceptable trade rather than reintroducing the mirrored design's
    server-admin-equivalent-everywhere problem.
+4. **A human can get a first login on a no-OIDC, operator-managed install —
+   without ever touching `AdminOnly` or `/api/admin/settings`.** This was
+   filed as `terdut-server#23` ("no API path to create a human login after
+   bootstrap") and diagnosed, at the time, as this server needing to let a
+   service account through `AdminOnly`. It doesn't: the fix lives entirely
+   in `terdut-operator`, because a team-scoped credential was *already*
+   owner-equivalent for `POST /api/teams/{teamID}/invites`, and invite
+   redemption (`POST /api/signup` with an `invite` token) bypasses
+   `signup_mode` entirely — `terdut-operator` just never grew a feature to
+   use either fact. Its `TerdutTeam` controller now mints and surfaces one
+   via its own existing team-scoped credential (`spec.invite`,
+   `status.inviteSecretRef`, see that repo's own docs), so a human joins a
+   CRD-managed team by a real invite link, the same way anyone else would.
+   `terdut-server#23` is closed with this note once that feature ships — its
+   named routes stay human-only, correctly, not a gap.
 
 ## Suggested sequencing
 

TEAM-LOOKUP.md

@@ -0,0 +1,73 @@
+# Team lookup for service accounts: closing terdut-operator's create-path crash window
+
+This is a design note for a feature, not an implementation plan — same posture as
+`SERVICE-ACCOUNTS.md`, and raised for the same reason: `terdut-operator`'s `TerdutTeam`
+controller (ROADMAP.md Stage 2, a separate repo, no shared code) hit a gap this server has
+no answer for yet.
+
+## The problem
+
+`POST /api/teams` (`handleCreateTeam`, confirmed against `internal/api/teams.go`) lets an
+instance-scoped service account create a team — it has its own explicit
+`isInstanceServiceAccount(...)` branch alongside the human-user path, not gated by
+`AdminOnly`. If that call succeeds server-side but the caller (`TerdutTeam`'s controller)
+crashes before persisting the resulting team ID locally, a retry's `POST` 409s on the name's
+unique constraint (confirmed: the `isUniqueViolation` branch in the same handler).
+
+Recovering from that 409 means looking the team up by name, and nothing today permits that
+for a service account:
+
+- `GET /api/teams` (`handleListTeams`) answers "what teams does the *caller* belong to", via
+  a `team_members` join keyed on `userFromContext`'s `caller.ID` — confirmed against source.
+  A service account is never a member of anything, so this always returns empty for one,
+  regardless of what exists.
+- `GET /api/admin/teams` (`handleAdminListTeams`) is gated by `AdminOnly`, and `AdminOnly`'s
+  actual code (`internal/api/middleware.go`) checks only `userFromContext(...).IsAdmin` — no
+  branch for a service account at all, confirmed against source. This contradicts
+  `SERVICE-ACCOUNTS.md`'s own text, which claims "an instance-scoped [service account
+  satisfies] `AdminOnly` for team-creation/listing purposes" — that claim doesn't match this
+  endpoint's actual, shipped code. (Team *creation* is fine: `handleCreateTeam` isn't behind
+  `AdminOnly` at all, it has its own check. Only the listing half of that sentence is wrong.)
+
+This is exactly the shape of gap `SERVICE-ACCOUNTS.md`'s own `GET /api/service-accounts?name=`
+closed for service accounts themselves (confirmed: that endpoint's own comment —
+"the name lookup is open to any authenticated caller... what lets a service account find its
+own account on the 403 that follows a second POST"). Teams never got the equivalent, because
+nothing needed it until an operator started creating them unattended.
+
+## Goals
+
+- A service-account-accessible way to look up one team by exact name, mirroring
+  `GET /api/service-accounts?name=` as closely as possible — same shape, same reasoning,
+  same low sensitivity of what it discloses.
+- No change to today's behavior for an empty/no-name request.
+
+## Proposed shape
+
+Extend `GET /api/teams` itself, the same way `handleListServiceAccounts` already branches on
+a `?name=` query param, rather than adding a new route:
+
+- `name` unset (today's behavior, unchanged): the caller's own teams, via `team_members`.
+- `name=<value>` set: look up that one team by exact name — a one-or-zero-length array, not
+  an error on no match, mirroring `GET /api/service-accounts?name=`'s own response shape and
+  status codes exactly. Deliberately **not** gated by `isInstanceServiceAccount` or
+  `AdminOnly`: a human caller who's already a member sees this same information in their own
+  team list regardless, and a non-member learning only that a name is taken — not who's in
+  the team, not any of its data — is the same low-sensitivity disclosure
+  `GET /api/service-accounts?name=` already accepts for service-account names.
+
+## What this unblocks
+
+Directly resolves the crash-window gap in `terdut-operator`'s `TerdutTeam` controller: on a
+409 from `POST /api/teams`, `GET /api/teams?name=<the same name>` — authenticated with the
+same instance-scoped credential that just got the 409 — finds the id, and the controller
+proceeds as if its own create had returned it directly. The same adopt-on-409 pattern already
+proven for service accounts (that repo's `DESIGN.md` §6 point 1, §5's general rule), not a
+new one.
+
+## Suggested sequencing
+
+Land this before `TerdutTeam`'s create path is implemented — the same reasoning
+`SERVICE-ACCOUNTS.md` gave for its own sequencing: writing that code against today's gap as a
+"known-temporary workaround" is wasted effort when the fix is this small and this
+well-precedented.

charts/terdut-server/Chart.yaml

@@ -15,5 +15,5 @@ type: application
 # appVersion and image.tag in values.yaml no longer agree, and that is not an oversight:
 # image.tag stays "latest", which is what a local install actually pulls. appVersion is
 # metadata and drives nothing.
-version: 0.33.0
-appVersion: "v0.33.0"
+version: 0.37.0
+appVersion: "v0.37.0"

charts/terdut-server/templates/deployment.yaml

@@ -6,21 +6,42 @@ metadata:
   labels:
     {{- include "terdut-server.labels" . | nindent 4 }}
 spec:
-  replicas: 1
+  replicas: {{ .Values.replicaCount }}
   selector:
     matchLabels:
       {{- include "terdut-server.selectorLabels" . | nindent 6 }}
-  # Recreate, not RollingUpdate, even though the PVC that forced it is gone: the
-  # sweeper and the notifier are unsynchronised singletons, and two replicas
-  # overlapping during a rollout would both page for the same incident.
+  # RollingUpdate, not Recreate: the sweeper, notifier and migration runner
+  # each take a Postgres advisory lock around their own pass, and new-incident
+  # creation on the first webhook for a brand-new groupKey resolves its own
+  # insert conflict -- so two replicas overlapping during a rollout no longer
+  # double-page, race a migration, or drop a webhook payload (v0.36.0). No
+  # explicit maxUnavailable/maxSurge: the 25%/25% default rounds to 0/1 at
+  # replicaCount: 2, which is zero-downtime already.
   strategy:
-    type: Recreate
+    type: RollingUpdate
   template:
     metadata:
       labels:
         {{- include "terdut-server.selectorLabels" . | nindent 8 }}
     spec:
       enableServiceLinks: false
+      {{- if .Values.database.waitForPostgres.enabled }}
+      initContainers:
+        - name: wait-for-postgres
+          image: "{{ .Values.database.waitForPostgres.image.repository }}:{{ .Values.database.waitForPostgres.image.tag }}"
+          imagePullPolicy: {{ .Values.database.waitForPostgres.image.pullPolicy }}
+          env:
+            - name: TERDUT_DB_DSN
+              value: {{ required "database.dsn is required" .Values.database.dsn | quote }}
+          command:
+            - sh
+            - -c
+            - |
+              until pg_isready -d "$TERDUT_DB_DSN"; do
+                echo "wait-for-postgres: not ready yet, retrying in 2s"
+                sleep 2
+              done
+      {{- end }}
       containers:
         - name: terdut-server
           image: "{{ .Values.image.repository }}:{{ .Values.image.tag }}"

charts/terdut-server/values.yaml

@@ -1,3 +1,10 @@
+# Safe above 1 since v0.36.0: the sweeper, notifier and migration runner each
+# take a Postgres advisory lock around their own pass, and a webhook that
+# loses the race to open a brand-new incident attaches to the winner's row
+# instead of dropping its payload. An image older than v0.36.0 does not have
+# these guards -- do not raise this against one.
+replicaCount: 2
+
 networking:
   hostname: "terdut.example.com"
   servicePort: 8080
@@ -33,6 +40,26 @@ database:
   passwordSecret:
     name: ""
     key: password
+  # Blocks the main container from starting until Postgres accepts
+  # connections. Without this, a Deployment created before Postgres has
+  # finished its very first boot -- initdb plus Patroni leader election, on a
+  # from-scratch postgres-operator cluster -- crash-loops a few times: the
+  # app's own ping-retry budget on startup (pingAttempts/pingRetryDelay in
+  # internal/db/db.go) is sized for a much shorter, different race --
+  # NetworkPolicy propagation, a few seconds -- not for genuine first-time
+  # cluster creation, which routinely takes longer, so it exhausts and the
+  # process exits before ever binding its HTTP port. A startupProbe cannot
+  # help here: the crash happens before there is anything to probe.
+  #
+  # pg_isready needs no credentials -- it reports PQPING_OK on anything that
+  # amounts to "a Postgres backend answered", including an auth challenge --
+  # so no PGPASSWORD is wired into this container.
+  waitForPostgres:
+    enabled: true
+    image:
+      repository: postgres
+      tag: "17-alpine"
+      pullPolicy: IfNotPresent
 
 service:
   type: ClusterIP

internal/api/advisory_lock_test.go

@@ -0,0 +1,122 @@
+package api
+
+// This file is internal (package api, not api_test) because withAdvisoryLock is
+// unexported and these tests exercise its locking semantics directly rather than
+// through the full StartArchiver/StartNotifier loop, which would make the "does
+// not run while held" case timing-dependent instead of deterministic. It opens a
+// plain connection to TERDUT_TEST_DSN rather than reusing testdb_test.go's
+// newTestDB, since that helper lives in the separate, already-compiled
+// api_test package and a Postgres advisory lock needs no schema or migration
+// to exercise.
+
+import (
+	"context"
+	"database/sql"
+	"os"
+	"testing"
+
+	_ "github.com/jackc/pgx/v5/stdlib"
+)
+
+// advisoryTestDB opens a plain, unmigrated connection to the test database. An
+// unset DSN fails rather than skips, matching testdb_test.go's rationale: a
+// suite that quietly tests nothing is worse than one that does not run.
+func advisoryTestDB(t *testing.T) *sql.DB {
+	t.Helper()
+
+	dsn := os.Getenv("TERDUT_TEST_DSN")
+	if dsn == "" {
+		t.Fatalf("TERDUT_TEST_DSN is not set: these tests need Postgres.\n" +
+			"Run `make test-db` for a local one, then\n" +
+			"  export TERDUT_TEST_DSN=postgres://terdut:terdut@localhost:5432/terdut_test?sslmode=disable")
+	}
+
+	db, err := sql.Open("pgx", dsn)
+	if err != nil {
+		t.Fatalf("connect to TERDUT_TEST_DSN: %v", err)
+	}
+	t.Cleanup(func() { db.Close() })
+	return db
+}
+
+func TestWithAdvisoryLock_RunsWhenFree(t *testing.T) {
+	db := advisoryTestDB(t)
+	ctx := context.Background()
+
+	ran := false
+	withAdvisoryLock(ctx, db, archiverLockKey, "test", func() { ran = true })
+
+	if !ran {
+		t.Fatal("fn did not run although the lock was free")
+	}
+}
+
+func TestWithAdvisoryLock_SkipsWhileHeldElsewhere(t *testing.T) {
+	db := advisoryTestDB(t)
+	ctx := context.Background()
+
+	// Hold the lock on a connection of our own, standing in for another
+	// replica mid-pass.
+	holder, err := db.Conn(ctx)
+	if err != nil {
+		t.Fatalf("acquire holder connection: %v", err)
+	}
+	defer holder.Close()
+	if _, err := holder.ExecContext(ctx, "SELECT pg_advisory_lock($1)", archiverLockKey); err != nil {
+		t.Fatalf("pre-acquire lock: %v", err)
+	}
+
+	ran := false
+	withAdvisoryLock(ctx, db, archiverLockKey, "test", func() { ran = true })
+	if ran {
+		t.Fatal("fn ran although another connection already held the lock")
+	}
+
+	if _, err := holder.ExecContext(ctx, "SELECT pg_advisory_unlock($1)", archiverLockKey); err != nil {
+		t.Fatalf("release held lock: %v", err)
+	}
+
+	// Now that the holder released it, the next caller should get it.
+	ran = false
+	withAdvisoryLock(ctx, db, archiverLockKey, "test", func() { ran = true })
+	if !ran {
+		t.Fatal("fn did not run after the other connection released the lock")
+	}
+}
+
+func TestWithAdvisoryLock_ReleasesAfterFnReturns(t *testing.T) {
+	db := advisoryTestDB(t)
+	ctx := context.Background()
+
+	withAdvisoryLock(ctx, db, notifierLockKey, "test", func() {})
+
+	// If the first call had leaked the lock, this one would see it held and
+	// skip, leaving ran false.
+	ran := false
+	withAdvisoryLock(ctx, db, notifierLockKey, "test", func() { ran = true })
+	if !ran {
+		t.Fatal("fn did not run on a later call: the earlier call leaked its lock")
+	}
+}
+
+func TestWithAdvisoryLock_KeysAreIndependent(t *testing.T) {
+	db := advisoryTestDB(t)
+	ctx := context.Background()
+
+	holder, err := db.Conn(ctx)
+	if err != nil {
+		t.Fatalf("acquire holder connection: %v", err)
+	}
+	defer holder.Close()
+	if _, err := holder.ExecContext(ctx, "SELECT pg_advisory_lock($1)", archiverLockKey); err != nil {
+		t.Fatalf("pre-acquire archiver lock: %v", err)
+	}
+	defer holder.ExecContext(ctx, "SELECT pg_advisory_unlock($1)", archiverLockKey)
+
+	// Holding archiverLockKey must not block notifierLockKey.
+	ran := false
+	withAdvisoryLock(ctx, db, notifierLockKey, "test", func() { ran = true })
+	if !ran {
+		t.Fatal("fn did not run under a different key although only archiverLockKey was held")
+	}
+}

internal/api/alertmanager.go

@@ -376,6 +376,17 @@ func incidentForGroup(ctx context.Context, tx *sql.Tx, notify NotifyConfig, team
 // its own. Hence the querier rather than a *sql.Tx. A nil severity leaves the
 // column for refreshSeverity to fill from the member alerts; the sweeper passes
 // one because its incidents have no members to derive it from.
+//
+// Both callers get here only after their own SELECT found no open incident for
+// this group_key — but on more than one replica, two webhook deliveries for the
+// very first occurrence of a brand-new group_key can both pass that SELECT
+// before either INSERTs. ON CONFLICT DO NOTHING against
+// incidents_open_group_key_idx is what makes the loser's INSERT a no-op instead
+// of a unique-violation error that would otherwise roll back its entire
+// payload; existingOpenIncident then hands it the winner's row. Postgres
+// resolves that conflict only once the winner's transaction has committed (or
+// rolled back), so by the time this RETURNING comes back empty, the winner's
+// row is guaranteed visible to that follow-up SELECT.
 func openIncident(ctx context.Context, q querier, notify NotifyConfig, teamID int64, groupKey, title string, groupLabels map[string]string, severity *string) (int64, error) {
 	onCall, err := currentOnCall(ctx, q, teamID)
 	if err != nil {
@@ -391,10 +402,18 @@ func openIncident(ctx context.Context, q querier, notify NotifyConfig, teamID in
 	err = q.QueryRowContext(ctx, `
 		INSERT INTO incidents (team_id, group_key, title, group_labels, signature, status, severity, triggered_at, assigned_to)
 		VALUES ($1, $2, $3, $4::jsonb, $5, 'triggered', $6, $7, $8)
+		ON CONFLICT (team_id, group_key) WHERE resolved_at IS NULL DO NOTHING
 		RETURNING id`,
 		teamID, groupKey, title, string(labelsJSON), incidentSignature(groupLabels, title), severity,
 		time.Now().Unix(), onCall).Scan(&id)
-	if err != nil {
+	switch {
+	case err == sql.ErrNoRows:
+		// Lost the race: someone else's incident for this group_key exists now.
+		// Everything below — the trigger event, assignment, page, escalation
+		// clock — already happened for that row when it was created; attach to
+		// it rather than fail this call (and the whole payload) outright.
+		return existingOpenIncident(ctx, q, teamID, groupKey)
+	case err != nil:
 		return 0, err
 	}
 
@@ -423,6 +442,21 @@ func openIncident(ctx context.Context, q querier, notify NotifyConfig, teamID in
 	return id, nil
 }
 
+// existingOpenIncident looks up the open incident openIncident's own INSERT just
+// lost a conflict against — the same lookup incidentForGroup does before ever
+// calling openIncident, repeated here for the caller that arrived second.
+func existingOpenIncident(ctx context.Context, q querier, teamID int64, groupKey string) (int64, error) {
+	var id int64
+	err := q.QueryRowContext(ctx,
+		"SELECT id FROM incidents WHERE team_id = $1 AND group_key = $2 AND resolved_at IS NULL",
+		teamID, groupKey,
+	).Scan(&id)
+	if err != nil {
+		return 0, err
+	}
+	return id, nil
+}
+
 // linkAlert adds an alert to an incident, emitting a timeline entry only the
 // first time. Re-sends of an already-linked alert are silent.
 func linkAlert(ctx context.Context, tx *sql.Tx, incidentID, alertID int64) error {

internal/api/alertmanager_test.go

@@ -0,0 +1,112 @@
+package api_test
+
+import (
+	"bytes"
+	"encoding/json"
+	"fmt"
+	"net/http"
+	"sync"
+	"testing"
+)
+
+// TestWebhook_ConcurrentFirstOccurrenceOpensOneIncident reproduces two
+// replicas racing the very first webhook delivery for a brand-new group_key:
+// both see no open incident yet (incidentForGroup's own SELECT finds
+// nothing) and race openIncident's INSERT.
+//
+// The DB's own unique index already guarantees at most one incident either
+// way, with or without this fix — so "exactly one incident" alone cannot
+// tell a fixed run from a broken one. What ON CONFLICT handling actually
+// changes is what happens to the *loser*: before it, the loser's INSERT hit
+// incidents_open_group_key_idx's unique violation, which — since
+// upsertAlerts ran earlier in that same transaction — rolled back its whole
+// payload, alert insert included. ingest's error is only logged and
+// receiveWebhook answers 200 regardless, so nothing ever retried it: the
+// loser's alert silently never existed. That is the regression signal this
+// test checks — every caller's fingerprint must show up in /api/alerts, not
+// just the winner's.
+func TestWebhook_ConcurrentFirstOccurrenceOpensOneIncident(t *testing.T) {
+	s := newTS(t)
+
+	const callers = 8
+	const groupKey = "race-group"
+
+	// Every caller needs its own fingerprint. A shared one would serialize all
+	// of them at upsertAlerts' own ON CONFLICT (team_id, fingerprint) row lock,
+	// long before any of them reached incidentForGroup — which would hide the
+	// very race this test exists to force.
+	bodies := make([][]byte, callers)
+	for i := range callers {
+		payload := map[string]any{
+			"version":     "4",
+			"status":      "firing",
+			"groupKey":    groupKey,
+			"groupLabels": map[string]string{"alertname": "RaceAlert"},
+			"alerts": []map[string]any{amAlert(fmt.Sprintf("fp-race-%d", i), "RaceAlert", "firing",
+				"2026-05-20T10:00:00Z", "0001-01-01T00:00:00Z", nil)},
+		}
+		bodies[i], _ = json.Marshal(payload)
+	}
+
+	// A start line, so every request is fired as close to simultaneously as
+	// goroutine scheduling allows, rather than trickling out one dial at a
+	// time — the race window is the gap between incidentForGroup's SELECT and
+	// openIncident's INSERT, which a staggered start could easily miss.
+	var ready sync.WaitGroup
+	start := make(chan struct{})
+	statuses := make([]int, callers)
+	var wg sync.WaitGroup
+	for i := range callers {
+		ready.Add(1)
+		wg.Add(1)
+		go func(i int) {
+			defer wg.Done()
+			ready.Done()
+			<-start
+			resp, err := http.Post(s.URL+"/api/integrations/"+s.ingestKey+"/alertmanager",
+				"application/json", bytes.NewReader(bodies[i]))
+			if err != nil {
+				t.Errorf("post webhook #%d: %v", i, err)
+				return
+			}
+			defer resp.Body.Close()
+			statuses[i] = resp.StatusCode
+		}(i)
+	}
+	ready.Wait()
+	close(start)
+	wg.Wait()
+
+	for i, code := range statuses {
+		if code != http.StatusOK {
+			t.Errorf("webhook #%d returned %d, want 200", i, code)
+		}
+	}
+
+	var matched []any
+	for _, inc := range listIncidents(t, s, "") {
+		if inc["group_key"] == groupKey {
+			matched = append(matched, inc["id"])
+		}
+	}
+	if len(matched) != 1 {
+		t.Fatalf("expected exactly 1 incident for group_key %q after %d concurrent deliveries, got %d: %v",
+			groupKey, callers, len(matched), matched)
+	}
+
+	var alerts []map[string]any
+	decode(t, s.req(t, http.MethodGet, "/api/alerts", nil), &alerts)
+	seen := map[string]bool{}
+	for _, a := range alerts {
+		if fp, ok := a["fingerprint"].(string); ok {
+			seen[fp] = true
+		}
+	}
+	for i := range callers {
+		fp := fmt.Sprintf("fp-race-%d", i)
+		if !seen[fp] {
+			t.Errorf("alert %q is missing: its delivery's whole payload was silently rolled back "+
+				"when it lost the race for the incident", fp)
+		}
+	}
+}

internal/api/archiver.go

@@ -15,6 +15,12 @@ const (
 	// expiryGrace absorbs clock skew and notification latency before an alert
 	// whose ends_at watermark has passed is treated as stale.
 	expiryGrace = 5 * time.Minute
+
+	// archiverLockKey is the Postgres advisory lock the sweeper takes for the
+	// duration of each pass, so that running more than one replica does not run
+	// the sweep concurrently on all of them. Its value has no meaning beyond
+	// being distinct from notifierLockKey.
+	archiverLockKey int64 = 7265_0001
 )
 
 // StartArchiver runs the alert sweeper until ctx is cancelled, starting with an
@@ -23,15 +29,25 @@ const (
 // the fallback, not the setting: each pass reads the current value from the
 // settings table, so an administrator's change takes effect on the next tick
 // instead of at the next restart.
+//
+// Each pass runs under archiverLockKey (see withAdvisoryLock), so that on more
+// than one replica only whichever instance's tick takes the lock first actually
+// sweeps; the rest skip that tick rather than racing the same pass.
 func StartArchiver(ctx context.Context, db *sql.DB, archiveAfter, staleAfter time.Duration, notify NotifyConfig) {
 	ticker := time.NewTicker(sweepInterval)
 	defer ticker.Stop()
 
-	Sweep(ctx, db, archiveAfter, staleAfter, notify)
+	sweep := func() {
+		withAdvisoryLock(ctx, db, archiverLockKey, "sweeper", func() {
+			Sweep(ctx, db, archiveAfter, staleAfter, notify)
+		})
+	}
+
+	sweep()
 	for {
 		select {
 		case <-ticker.C:
-			Sweep(ctx, db, archiveAfter, staleAfter, notify)
+			sweep()
 		case <-ctx.Done():
 			return
 		}

internal/api/auth.go

@@ -279,7 +279,11 @@ type meResponse struct {
 // between the login form and the app, since it cannot read its own cookie.
 func handleMe(db *sql.DB) http.HandlerFunc {
 	return func(w http.ResponseWriter, r *http.Request) {
-		caller, _ := userFromContext(r.Context())
+		caller, ok := userFromContext(r.Context())
+		if !ok {
+			respond(w, http.StatusForbidden, errResp("this endpoint is for human accounts only"))
+			return
+		}
 		user, err := fetchUser(r.Context(), db, caller.ID)
 		if err != nil {
 			respond(w, http.StatusInternalServerError, errResp("internal error"))

internal/api/caller.go

@@ -0,0 +1,123 @@
+package api
+
+import (
+	"context"
+	"fmt"
+
+	"git.ryuvia.com/niklas/terdut-server/internal/models"
+)
+
+// Caller is the one principal type every authorization predicate in this
+// package reads from. Before this, a human (ctxUser + ctxTeams) and a
+// service account (ctxServiceAccount + a synthetic ctxTeams entry) were two
+// parallel, un-unified context representations — every predicate had to
+// remember which one(s) it needed to check, and the ones that forgot either
+// 403'd a service account that should have been let through (terdut-server#23,
+// terdut-operator#3), crashed on an unchecked zero-value user ID (handleMe,
+// handleTestNotification), or silently no-op'd (handleDismissOnboarding).
+// serveAs and serveAsServiceAccount now both build exactly one Caller and
+// store it under one context key; everything else in this file is a read
+// of one of its methods.
+type Caller struct {
+	// user is set for a human caller (session cookie or a user's own API
+	// key), nil for a service account of either scope.
+	user *models.User
+
+	// sa is set for a service-account caller, nil for a human.
+	sa *serviceAccountPrincipal
+
+	// memberships is the caller's real team_members rows for a human, or —
+	// for a team-scoped service account — the single synthetic owner
+	// membership serveAsServiceAccount injects (see its own comment for
+	// why). Always nil for an instance-scoped service account: it acts on
+	// teams by id, not by belonging to one.
+	memberships []membership
+}
+
+// AsHuman returns the real user behind this caller, or false for a service
+// account of either scope. Every handler that needs a real user_id to act
+// on behalf of — not just "is this caller sufficiently privileged" — calls
+// this and handles the false case explicitly, replacing the unchecked
+// userFromContext(ctx) zero-value reads that used to silently misbehave for
+// a service-account caller.
+func (c Caller) AsHuman() (models.User, bool) {
+	if c.user == nil {
+		return models.User{}, false
+	}
+	return *c.user, true
+}
+
+// IsAdmin is true only for a human system administrator — never for a
+// service account, of either scope, under any circumstance. AdminOnly and
+// requireSelfOrAdmin key on this and nothing else: user management and
+// /api/admin/settings stay human-only forever, by design (SERVICE-ACCOUNTS.md).
+func (c Caller) IsAdmin() bool {
+	return c.user != nil && c.user.IsAdmin
+}
+
+// IsInstanceServiceAccount reports whether this caller is specifically an
+// instance-scoped service account — never true for a human, including a
+// human admin. handleCreateTeam needs exactly this: a human creates a team
+// by being a human (and becomes its owner as a side effect), an
+// instance-scoped service account creates one with no human owner at all;
+// the two paths are not interchangeable, so this predicate must not also
+// admit a human admin the way MayActAsInstanceAdmin deliberately does.
+func (c Caller) IsInstanceServiceAccount() bool {
+	return c.sa != nil && c.sa.scope == models.ServiceAccountScopeInstance
+}
+
+// Role reports the caller's role in teamID, and whether they belong to it
+// at all.
+func (c Caller) Role(teamID int64) (string, bool) {
+	for _, m := range c.memberships {
+		if m.teamID == teamID {
+			return m.role, true
+		}
+	}
+	return "", false
+}
+
+// TeamIDs lists every team this caller belongs to: a human's real
+// memberships, or a team-scoped service account's own single team. Always
+// empty for an instance-scoped service account.
+func (c Caller) TeamIDs() []int64 {
+	ids := make([]int64, 0, len(c.memberships))
+	for _, m := range c.memberships {
+		ids = append(ids, m.teamID)
+	}
+	return ids
+}
+
+// ServiceAccountID reports this caller's own service-account id, for the
+// "may manage/rotate its own credential" self-check in
+// callerMayManageServiceAccount, and for OperatorModeBlock's "any service
+// account passes" rule.
+func (c Caller) ServiceAccountID() (int64, bool) {
+	if c.sa == nil {
+		return 0, false
+	}
+	return c.sa.id, true
+}
+
+// Identity is a stable, log/audit-facing string distinguishing a human
+// caller from a service account — "user:42" or "service-account:7". Not
+// wired into any database column today (incidents.go's acknowledged_by/
+// assigned_to/user_id are explicitly out of scope for this change — that
+// needs its own schema migration, tracked separately), but this is the one
+// place in the request path that already knows which kind of caller this
+// is, and that follow-up will want exactly this accessor.
+func (c Caller) Identity() string {
+	switch {
+	case c.user != nil:
+		return fmt.Sprintf("user:%d", c.user.ID)
+	case c.sa != nil:
+		return fmt.Sprintf("service-account:%d", c.sa.id)
+	default:
+		return "unknown"
+	}
+}
+
+func callerFromContext(ctx context.Context) (Caller, bool) {
+	c, ok := ctx.Value(ctxCaller).(Caller)
+	return c, ok
+}

internal/api/helpers.go

@@ -1,8 +1,11 @@
 package api
 
 import (
+	"context"
+	"database/sql"
 	"encoding/json"
 	"errors"
+	"log"
 	"net/http"
 	"strconv"
 	"strings"
@@ -77,3 +80,39 @@ func decodeJSON(r *http.Request, v any) error {
 func errResp(msg string) map[string]string {
 	return map[string]string{"error": msg}
 }
+
+// withAdvisoryLock runs fn only if it can take the named Postgres advisory lock on a
+// dedicated connection, and skips fn otherwise. This is what keeps the archiver and
+// notifier safe to run on more than one replica: whichever instance's tick gets there
+// first does the work; the rest see the lock held and simply wait for their next tick
+// instead of running the same pass concurrently.
+//
+// pg_try_advisory_lock is session-scoped, so taking and releasing it must happen on the
+// same connection, reserved via db.Conn rather than borrowed from the pool's shared
+// connections fn itself may use — and released (unlocked, then closed) before returning,
+// since a session lock otherwise outlives this call and leaks onto whatever reuses the
+// pooled connection next.
+func withAdvisoryLock(ctx context.Context, db *sql.DB, key int64, name string, fn func()) {
+	conn, err := db.Conn(ctx)
+	if err != nil {
+		log.Printf("%s: advisory lock: acquire connection: %v", name, err)
+		return
+	}
+	defer conn.Close()
+
+	var locked bool
+	if err := conn.QueryRowContext(ctx, "SELECT pg_try_advisory_lock($1)", key).Scan(&locked); err != nil {
+		log.Printf("%s: advisory lock: %v", name, err)
+		return
+	}
+	if !locked {
+		return // another replica is already running this pass
+	}
+	defer func() {
+		if _, err := conn.ExecContext(ctx, "SELECT pg_advisory_unlock($1)", key); err != nil {
+			log.Printf("%s: advisory unlock: %v", name, err)
+		}
+	}()
+
+	fn()
+}

internal/api/incident_store.go

@@ -261,14 +261,17 @@ func resolveIfSettled(ctx context.Context, q querier, incidentID int64) (bool, e
 }
 
 // acknowledgeIncident records that userID has picked an incident up, and reports
-// whether it changed anything — an already-resolved incident is left alone.
-// Shared by the authenticated handler and the Acknowledge button in a push
-// notification, so both write the same state and the same timeline entry.
+// whether it changed anything — an already-resolved or already-acknowledged
+// incident is left alone, so a second acknowledge (a retried request, or a
+// stale push notification tapped after the web UI already acked it) is a
+// no-op rather than a second "acknowledged" timeline entry. Shared by the
+// authenticated handler and the Acknowledge button in a push notification,
+// so both write the same state and the same timeline entry.
 func acknowledgeIncident(ctx context.Context, q querier, incidentID, userID int64) (bool, error) {
 	res, err := q.ExecContext(ctx, `
 		UPDATE incidents
 		   SET status = 'acknowledged', acknowledged_by = $1, acknowledged_at = $2
-		 WHERE id = $3 AND resolved_at IS NULL`,
+		 WHERE id = $3 AND status = 'triggered'`,
 		userID, time.Now().Unix(), incidentID)
 	if err != nil {
 		return false, err

internal/api/incidents.go

@@ -197,10 +197,20 @@ func handleIncidentAcknowledge(db *sql.DB) http.HandlerFunc {
 			return
 		}
 		if !acked {
-			if !incidentExists(w, r, db, id) {
+			// incidentIDParam above already confirmed the incident exists, so this
+			// is either resolved, or already acknowledged — the latter is now a
+			// no-op rather than an error, since the caller's desired state
+			// (acknowledged) already holds.
+			inc, err := fetchIncident(r.Context(), db, id)
+			if err != nil {
+				respond(w, http.StatusInternalServerError, errResp("internal error"))
 				return
 			}
-			respond(w, http.StatusConflict, errResp("incident is resolved"))
+			if inc.Status == "resolved" {
+				respond(w, http.StatusConflict, errResp("incident is resolved"))
+				return
+			}
+			respond(w, http.StatusOK, inc)
 			return
 		}
 		respondIncident(w, r, db, id)

internal/api/incidents_test.go

@@ -338,6 +338,48 @@ func TestIncident_Acknowledge(t *testing.T) {
 	}
 }
 
+// A second acknowledge — a retried request, or a stale push notification
+// tapped after the web UI already acked it — must be a no-op: same state,
+// no second "acknowledged" timeline entry. Regression test for the bug
+// described in issue #26 ("two acknowledged entries look like a bug").
+func TestIncident_AcknowledgeTwiceIsIdempotent(t *testing.T) {
+	s := newTS(t)
+	postWebhook(t, s, []map[string]any{
+		amAlert("fp-ack2", "Y", "firing", "2026-05-20T10:00:00Z", zeroTime, nil),
+	})
+
+	resp := s.req(t, http.MethodPost, "/api/incidents/1/acknowledge", nil)
+	if resp.StatusCode != http.StatusOK {
+		t.Fatalf("first acknowledge returned %d", resp.StatusCode)
+	}
+	var first map[string]any
+	decode(t, resp, &first)
+
+	resp = s.req(t, http.MethodPost, "/api/incidents/1/acknowledge", nil)
+	if resp.StatusCode != http.StatusOK {
+		t.Fatalf("second acknowledge returned %d, want 200 (idempotent)", resp.StatusCode)
+	}
+	var second map[string]any
+	decode(t, resp, &second)
+	if second["status"] != "acknowledged" {
+		t.Errorf("expected status still acknowledged, got %v", second["status"])
+	}
+	if second["acknowledged_by"] != first["acknowledged_by"] {
+		t.Errorf("expected the same acknowledged_by, got %v then %v", first["acknowledged_by"], second["acknowledged_by"])
+	}
+
+	types := eventTypes(timeline(t, s, 1))
+	n := 0
+	for _, ty := range types {
+		if ty == "acknowledged" {
+			n++
+		}
+	}
+	if n != 1 {
+		t.Errorf("expected exactly one acknowledged event, got %d in %v", n, types)
+	}
+}
+
 func TestIncident_ManualResolveIsTerminal(t *testing.T) {
 	s := newTS(t)
 	postWebhook(t, s, []map[string]any{

internal/api/middleware.go

@@ -16,10 +16,13 @@ import (
 type contextKey string
 
 const (
-	ctxUser           contextKey = "user"
-	ctxSession        contextKey = "session"
-	ctxTeams          contextKey = "teams"
-	ctxServiceAccount contextKey = "service_account"
+	// ctxCaller holds the one Caller (see caller.go) every authorization
+	// predicate in this package reads from — a human and a service account
+	// used to be two parallel, un-unified context keys (ctxUser/ctxTeams vs.
+	// ctxServiceAccount); this is why that was a mistake, not a smaller
+	// version of the same idea.
+	ctxCaller  contextKey = "caller"
+	ctxSession contextKey = "session"
 )
 
 // AuthMiddleware accepts either of the two credentials the server issues: an
@@ -179,8 +182,7 @@ func serveAs(w http.ResponseWriter, r *http.Request, next http.Handler, db *sql.
 		return
 	}
 
-	ctx := context.WithValue(r.Context(), ctxTeams, teams)
-	ctx = context.WithValue(ctx, ctxUser, u)
+	ctx := context.WithValue(r.Context(), ctxCaller, Caller{user: &u, memberships: teams})
 	if sessionID != 0 {
 		ctx = context.WithValue(ctx, ctxSession, sessionID)
 	}
@@ -192,9 +194,13 @@ func hashToken(token string) string {
 	return hex.EncodeToString(h[:])
 }
 
+// userFromContext is a thin compatibility wrapper over Caller.AsHuman(), so
+// every call site written before the Caller abstraction (alerts.go,
+// incidents.go, schedule.go, stats.go, and more) needs no change and keeps
+// its exact existing behavior.
 func userFromContext(ctx context.Context) (models.User, bool) {
-	u, ok := ctx.Value(ctxUser).(models.User)
-	return u, ok
+	c, _ := callerFromContext(ctx)
+	return c.AsHuman()
 }
 
 // serviceAccountPrincipal is a service account as resolved from its key:
@@ -244,26 +250,21 @@ func serviceAccountFor(ctx context.Context, db *sql.DB, token string) (serviceAc
 // key is only ever set by the client that holds it, never attached by a
 // browser to a request another site makes.
 func serveAsServiceAccount(w http.ResponseWriter, r *http.Request, next http.Handler, sa serviceAccountPrincipal) {
-	ctx := r.Context()
+	var memberships []membership
 	if sa.scope == models.ServiceAccountScopeTeam {
-		ctx = context.WithValue(ctx, ctxTeams, []membership{{teamID: sa.teamID, role: models.RoleOwner}})
+		memberships = []membership{{teamID: sa.teamID, role: models.RoleOwner}}
 	}
-	ctx = context.WithValue(ctx, ctxServiceAccount, sa)
+	ctx := context.WithValue(r.Context(), ctxCaller, Caller{sa: &sa, memberships: memberships})
 	next.ServeHTTP(w, r.WithContext(ctx))
 }
 
-func serviceAccountFromContext(ctx context.Context) (serviceAccountPrincipal, bool) {
-	sa, ok := ctx.Value(ctxServiceAccount).(serviceAccountPrincipal)
-	return sa, ok
-}
-
-// isInstanceServiceAccount reports whether the caller is an instance-scoped
-// service account — the one identity allowed to create a team and mint a
-// team-scoped account against any of them, the two things system
-// administration can already do that this extends to automation.
+// isInstanceServiceAccount is a thin compatibility wrapper over
+// Caller.IsInstanceServiceAccount(), for call sites outside this package's
+// core predicates (handleCreateTeam, handleCreateServiceAccount) that
+// needed this exact, narrow check before the Caller abstraction existed.
 func isInstanceServiceAccount(ctx context.Context) bool {
-	sa, ok := serviceAccountFromContext(ctx)
-	return ok && sa.scope == models.ServiceAccountScopeInstance
+	c, _ := callerFromContext(ctx)
+	return c.IsInstanceServiceAccount()
 }
 
 // operatorReason marks a write that operator mode refused as such, distinct
@@ -291,7 +292,8 @@ func OperatorModeBlock(cfg config.Config) func(http.Handler) http.Handler {
 			return next
 		}
 		return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-			if _, ok := serviceAccountFromContext(r.Context()); ok {
+			caller, _ := callerFromContext(r.Context())
+			if _, ok := caller.ServiceAccountID(); ok {
 				next.ServeHTTP(w, r)
 				return
 			}
@@ -333,24 +335,15 @@ func callerMemberships(ctx context.Context, db *sql.DB, userID int64) ([]members
 // administration is about accounts, not about reading other people's incidents,
 // and an admin who needs to see a team's queue can add themselves to it.
 func callerTeamIDs(ctx context.Context) []int64 {
-	ms, _ := ctx.Value(ctxTeams).([]membership)
-	ids := make([]int64, 0, len(ms))
-	for _, m := range ms {
-		ids = append(ids, m.teamID)
-	}
-	return ids
+	c, _ := callerFromContext(ctx)
+	return c.TeamIDs()
 }
 
 // callerRole reports the caller's role in one team, and whether they are in it
 // at all.
 func callerRole(ctx context.Context, teamID int64) (string, bool) {
-	ms, _ := ctx.Value(ctxTeams).([]membership)
-	for _, m := range ms {
-		if m.teamID == teamID {
-			return m.role, true
-		}
-	}
-	return "", false
+	c, _ := callerFromContext(ctx)
+	return c.Role(teamID)
 }
 
 // requireTeamMember answers the request and reports false unless the caller

internal/api/notifier.go

@@ -38,6 +38,13 @@ const (
 	ackTokenTTL = 24 * time.Hour
 )
 
+// notifierLockKey is the Postgres advisory lock the notifier takes for the
+// duration of each pass, so that running more than one replica does not
+// deliver (or double-deliver) the same notification from more than one of
+// them at once. Its value has no meaning beyond being distinct from
+// archiverLockKey.
+const notifierLockKey int64 = 7265_0002
+
 // Notification kinds, recording why a push was sent.
 const (
 	notifyTriggered = "triggered"
@@ -97,6 +104,10 @@ var notifyClient = &http.Client{Timeout: 10 * time.Second}
 
 // StartNotifier delivers queued notifications until ctx is cancelled, starting
 // with an immediate pass so a restart flushes whatever the last one left behind.
+//
+// Each pass runs under notifierLockKey (see withAdvisoryLock), so that on more
+// than one replica only whichever instance's tick takes the lock first actually
+// delivers; the rest skip that tick rather than racing the same pass.
 func StartNotifier(ctx context.Context, db *sql.DB, cfg NotifyConfig) {
 	if !cfg.enabled() {
 		log.Print("notifier: disabled (no ntfy URL configured)")
@@ -107,11 +118,17 @@ func StartNotifier(ctx context.Context, db *sql.DB, cfg NotifyConfig) {
 	ticker := time.NewTicker(notifyInterval)
 	defer ticker.Stop()
 
-	NotifySweep(ctx, db, cfg)
+	sweep := func() {
+		withAdvisoryLock(ctx, db, notifierLockKey, "notifier", func() {
+			NotifySweep(ctx, db, cfg)
+		})
+	}
+
+	sweep()
 	for {
 		select {
 		case <-ticker.C:
-			NotifySweep(ctx, db, cfg)
+			sweep()
 		case <-ctx.Done():
 			return
 		}

internal/api/notify_ack.go

@@ -66,12 +66,19 @@ func handleNotifyAck(db *sql.DB) http.HandlerFunc {
 			return
 		}
 		if !acked {
-			// The incident closed between the page and the tap. Nothing to do,
-			// and nothing the responder did wrong — report the state, not an error,
-			// so ntfy shows a success toast rather than a failure.
+			// Either the incident closed between the page and the tap, or it was
+			// already acknowledged (e.g. from the web UI, or an earlier tap of
+			// the same button) — either way nothing the responder did wrong, so
+			// report the actual state rather than assuming "resolved", and let
+			// ntfy show a success toast rather than a failure.
+			inc, err := fetchIncident(r.Context(), db, incidentID)
+			if err != nil {
+				respond(w, http.StatusInternalServerError, errResp("internal error"))
+				return
+			}
 			respond(w, http.StatusOK, map[string]any{
 				"incident_id": incidentID,
-				"status":      "resolved",
+				"status":      inc.Status,
 			})
 			return
 		}

internal/api/notify_test.go

@@ -408,6 +408,47 @@ func TestNotify_AckButtonAcknowledgesIncident(t *testing.T) {
 	}
 }
 
+// The ack token isn't single-use (it stays valid for a day, in case the
+// first tap never reaches the server), so tapping the same notification's
+// Acknowledge button twice is a real scenario, not just a retried request.
+// It must report the incident's actual state, not assume "resolved" —
+// see handleNotifyAck's !acked branch — and must not log a second
+// "acknowledged" event.
+func TestNotify_AckButtonTwiceIsIdempotent(t *testing.T) {
+	s, f := notifyTS(t, api.NotifyConfig{PublicURL: "https://terdut.example.com"})
+
+	fireCritical(t, s)
+	s.sweepNotify(t)
+
+	ackURL := f.messages()[0].Actions[0].URL
+	path := ackURL[strings.Index(ackURL, "/api/notify/ack/"):]
+
+	for i := range 2 {
+		resp, err := http.Post(s.URL+path, "application/json", nil)
+		if err != nil {
+			t.Fatalf("ack %d: %v", i+1, err)
+		}
+		var body map[string]any
+		decode(t, resp, &body)
+		if resp.StatusCode != http.StatusOK {
+			t.Fatalf("ack %d returned %d", i+1, resp.StatusCode)
+		}
+		if body["status"] != "acknowledged" {
+			t.Errorf("ack %d: expected status acknowledged, got %v", i+1, body["status"])
+		}
+	}
+
+	n := 0
+	for _, ty := range eventTypes(timeline(t, s, 1)) {
+		if ty == "acknowledged" {
+			n++
+		}
+	}
+	if n != 1 {
+		t.Errorf("expected exactly one acknowledged event after two taps, got %d", n)
+	}
+}
+
 func TestNotify_AckRejectsUnknownToken(t *testing.T) {
 	s, _ := notifyTS(t, api.NotifyConfig{PublicURL: "https://terdut.example.com"})
 	fireCritical(t, s)

internal/api/oidc.go

@@ -48,6 +48,17 @@ const (
 	ssoNoEmail       ssoError = "no_email"       // the provider sent no email address
 	ssoEmailConflict ssoError = "email_conflict" // a local account has this email and cannot be linked
 	ssoDisabled      ssoError = "disabled"       // the linked account is disabled
+	// ssoNotBootstrapped: this identity has no existing account, and no user
+	// exists on this install yet either -- creating one here would race
+	// POST /api/bootstrap for the one gitops-managed installs expect to win
+	// it (terdut-operator's own DESIGN.md §1, §6), which has no way to
+	// recover if it loses. The person sees this for at most as long as it
+	// takes whatever is bootstrapping this install to finish; signing in
+	// again afterward hits the ordinary first-sign-in path. Found by
+	// terdut-operator#1: nothing stopped an otherwise-ordinary OIDC sign-in
+	// from quietly winning this race against an operator that assumed it
+	// was the only caller.
+	ssoNotBootstrapped ssoError = "not_bootstrapped"
 )
 
 // handleAuthConfig says how this server can be signed in to, so the login form
@@ -366,6 +377,20 @@ func resolveSSOUser(ctx context.Context, tx *sql.Tx, cfg config.OIDC, id *oidc.I
 			return 0, ssoEmailConflict
 		}
 	case errors.Is(err, sql.ErrNoRows):
+		// Creating the very first user is /api/bootstrap's own job (same
+		// gate, same table: SELECT COUNT(*) FROM users in handleBootstrap).
+		// An identity nobody has linked yet, on an install with no users at
+		// all, is exactly the race terdut-operator#1 found: whoever gets
+		// here first wins a slot the other side has no way to recover from
+		// losing. Refusing it here costs an otherwise-ordinary sign-in
+		// nothing but a retry once bootstrap has actually run.
+		var userCount int
+		if err := tx.QueryRowContext(ctx, "SELECT COUNT(*) FROM users").Scan(&userCount); err != nil {
+			return 0, err
+		}
+		if userCount == 0 {
+			return 0, ssoNotBootstrapped
+		}
 		userID, err = createSSOUser(ctx, tx, id)
 		if err != nil {
 			return 0, err

internal/api/oidc_test.go

@@ -543,6 +543,45 @@ func TestSSO_DisabledUserIsRefused(t *testing.T) {
 	}
 }
 
+// TestSSO_FirstUserIsRefusedUntilBootstrap is terdut-operator#1: an
+// otherwise-ordinary OIDC sign-in against a brand-new, not-yet-bootstrapped
+// install must not be allowed to create the first user and win the race
+// POST /api/bootstrap expects to win uncontested. Built directly over
+// api.NewRouter rather than newSSOTS/newTS, both of which bootstrap before
+// a test body ever runs -- exactly the state this test needs to not have yet.
+func TestSSO_FirstUserIsRefusedUntilBootstrap(t *testing.T) {
+	idp := newFakeIdP(t)
+	database := newTestDB(t)
+	srv := httptest.NewServer(api.NewRouter(database, api.NotifyConfig{PublicURL: "http://terdut.test"}, ssoConfig(idp), "test"))
+	t.Cleanup(srv.Close)
+
+	first := newBrowser(t, srv.URL)
+	first.CheckRedirect = func(*http.Request, []*http.Request) error { return http.ErrUseLastResponse }
+	if loc := signInSSO(t, idp, first, alice); loc != "/?sso_error=not_bootstrapped" {
+		t.Fatalf("sent to %q, want not_bootstrapped", loc)
+	}
+
+	// Bootstrap the install for real, the way terdut-operator's own
+	// reconcileBootstrap does.
+	resp, err := http.Post(srv.URL+"/api/bootstrap", "application/json",
+		strings.NewReader(`{"username":"admin","email":"admin@test.com"}`))
+	if err != nil {
+		t.Fatal(err)
+	}
+	resp.Body.Close()
+	if resp.StatusCode != http.StatusCreated {
+		t.Fatalf("bootstrap: %d", resp.StatusCode)
+	}
+
+	// The same identity, signing in again, is this install's ordinary first
+	// SSO user now -- no longer refused.
+	second := newBrowser(t, srv.URL)
+	second.CheckRedirect = func(*http.Request, []*http.Request) error { return http.ErrUseLastResponse }
+	if loc := signInSSO(t, idp, second, alice); loc != "/" {
+		t.Errorf("sent to %q after bootstrap, want success", loc)
+	}
+}
+
 func TestSSO_NoEmailIsRefused(t *testing.T) {
 	idp := newFakeIdP(t)
 	s := newSSOTS(t, idp)

internal/api/service_accounts.go

@@ -41,10 +41,14 @@ func callerIsAdmin(ctx context.Context) bool {
 	return ok && u.IsAdmin
 }
 
-// callerOwnsTeam reports whether the caller is a human owner of teamID. Built
-// on callerRole/ctxTeams like requireTeamOwner, but without writing a
-// response: callers here need to combine it with other ways of being
-// allowed, not stop at the first no.
+// callerOwnsTeam reports whether the caller is owner-equivalent for teamID:
+// a human owner, or that team's own team-scoped service account (its single
+// synthetic membership, serveAsServiceAccount — ratified in
+// SERVICE-ACCOUNTS.md as intentional, not an accident: a team-scoped
+// credential is that team's owner's reach, full stop, membership and
+// invites included). Built on callerRole like requireTeamOwner, but without
+// writing a response: callers here need to combine it with other ways of
+// being allowed, not stop at the first no.
 func callerOwnsTeam(ctx context.Context, teamID int64) bool {
 	role, ok := callerRole(ctx, teamID)
 	return ok && role == models.RoleOwner
@@ -173,9 +177,24 @@ func fetchServiceAccount(ctx context.Context, db *sql.DB, id int64) (models.Serv
 
 // callerMayManageServiceAccount reports whether the caller may mint or revoke
 // a key on sa: a system administrator, that team-scoped account's own human
-// owner, or the account rotating its own credential — which is not a
-// privilege escalation, the same reasoning requireSelfOrAdmin already rests
-// on for a user's own API keys.
+// owner, the account rotating its own credential (not a privilege
+// escalation, the same reasoning requireSelfOrAdmin already rests on for a
+// user's own API keys) — or, new, an instance-scoped service account
+// managing any team-scoped account.
+//
+// That last branch closes terdut-operator#3: handleCreateServiceAccount
+// already lets an instance-scoped caller *create* a team-scoped account for
+// any team (the branch below it, isInstanceServiceAccount(ctx)) — this
+// account didn't have an equivalent reach to *adopt or rotate* one it
+// didn't just create in the same call, which is exactly the recovery path
+// terdut-operator's own documented crash-window handling depends on
+// (DESIGN.md §5's general adopt-on-conflict rule): a reconcile that creates
+// the account successfully but crashes before persisting its credential
+// locally retries into a 409, and without this branch the only available
+// recovery — minting a fresh key on the now-existing account — 403'd
+// forever, with no way out. Granting it here is not a new power: it
+// mirrors the create-time reach this scope already has, just extended to
+// the retry path DESIGN.md's own crash-window reasoning requires.
 func callerMayManageServiceAccount(ctx context.Context, sa models.ServiceAccount) bool {
 	if callerIsAdmin(ctx) {
 		return true
@@ -183,7 +202,11 @@ func callerMayManageServiceAccount(ctx context.Context, sa models.ServiceAccount
 	if sa.TeamID != nil && callerOwnsTeam(ctx, *sa.TeamID) {
 		return true
 	}
-	if self, ok := serviceAccountFromContext(ctx); ok && self.id == sa.ID {
+	caller, _ := callerFromContext(ctx)
+	if id, ok := caller.ServiceAccountID(); ok && id == sa.ID {
+		return true
+	}
+	if sa.TeamID != nil && caller.IsInstanceServiceAccount() {
 		return true
 	}
 	return false

internal/api/service_accounts_test.go

@@ -181,10 +181,136 @@ func TestServiceAccount_TeamScopeManagesItsResources(t *testing.T) {
 	resp2.Body.Close()
 }
 
+// Documents the capability already granted at create time (handleCreateServiceAccount's
+// own callerOwnsTeam branch) also applies here: a team-scoped account is that
+// team's owner's reach, membership and further accounts included, not just
+// the handful of endpoints exercised above. Kept, not restricted, for
+// symmetry with the now-ratified membership/invite capability below.
+func TestServiceAccount_TeamScopeCanMintAnotherAccountForItsOwnTeam(t *testing.T) {
+	s := newTS(t)
+	instanceKey := createServiceAccount(t, s, s.key, "terdut-operator", models.ServiceAccountScopeInstance, 0)
+	teamA := createTeamAs(t, s, instanceKey, "team-a")
+	keyA := createServiceAccount(t, s, instanceKey, "team-a-sa", models.ServiceAccountScopeTeam, teamA)
+
+	resp := s.reqAs(t, keyA, http.MethodPost, "/api/service-accounts",
+		map[string]any{"name": "team-a-sa-2", "scope": models.ServiceAccountScopeTeam, "team_id": teamA})
+	if resp.StatusCode != http.StatusCreated {
+		t.Errorf("team-scoped account minting another account for its own team: %d", resp.StatusCode)
+	}
+	resp.Body.Close()
+}
+
+// SERVICE-ACCOUNTS.md ratifies this explicitly: a team-scoped account is
+// owner-equivalent for every requireTeamOwner endpoint, membership and
+// invites included — terdut-operator's own invite-minting feature depends on
+// exactly this. No test exercised handleCreateInvite from a service account
+// before this change, and it would have 500'd (created_by written as a bare
+// zero value against a NOT-validated-but-FK'd column) rather than succeeded;
+// see the signup_test.go addition for that half.
+func TestServiceAccount_TeamScopeManagesItsOwnInvites(t *testing.T) {
+	s := newTS(t)
+	instanceKey := createServiceAccount(t, s, s.key, "terdut-operator", models.ServiceAccountScopeInstance, 0)
+	teamA := createTeamAs(t, s, instanceKey, "team-a")
+	keyA := createServiceAccount(t, s, instanceKey, "team-a-sa", models.ServiceAccountScopeTeam, teamA)
+
+	var invite struct {
+		ID int64 `json:"id"`
+	}
+	resp := s.reqAs(t, keyA, http.MethodPost, "/api/teams/"+id64(teamA)+"/invites", map[string]any{})
+	if resp.StatusCode != http.StatusCreated {
+		t.Fatalf("team-scoped account creating an invite: %d", resp.StatusCode)
+	}
+	decode(t, resp, &invite)
+
+	var list []map[string]any
+	decode(t, s.reqAs(t, keyA, http.MethodGet, "/api/teams/"+id64(teamA)+"/invites", nil), &list)
+	if len(list) != 1 {
+		t.Errorf("expected the invite to list back, got %d", len(list))
+	}
+
+	if resp := s.reqAs(t, keyA, http.MethodDelete,
+		"/api/teams/"+id64(teamA)+"/invites/"+id64(invite.ID), nil); resp.StatusCode != http.StatusNoContent {
+		t.Errorf("team-scoped account revoking its own invite: %d", resp.StatusCode)
+	} else {
+		resp.Body.Close()
+	}
+}
+
 // ---------------------------------------------------------------------------
 // Key rotation
 // ---------------------------------------------------------------------------
 
+// terdut-operator#3: an instance-scoped account is already trusted to CREATE
+// a team-scoped account for any team (handleCreateServiceAccount's own
+// isInstanceServiceAccount branch) — this pins that it is equally trusted to
+// manage/rotate a key on one that already exists and that it did not just
+// create in this call, which is the exact shape of terdut-operator's own
+// crash-window recovery (mint succeeds, a later step is interrupted before
+// persisting the credential locally, and the next reconcile retries into a
+// 409 then needs to mint a fresh key on the now-existing account). Before
+// this fix, the second POST .../keys below 403'd forever.
+func TestServiceAccount_InstanceScopeAdoptsAnExistingTeamScopedAccountsKey(t *testing.T) {
+	s := newTS(t)
+	instanceKey := createServiceAccount(t, s, s.key, "terdut-operator", models.ServiceAccountScopeInstance, 0)
+	teamA := createTeamAs(t, s, instanceKey, "team-a")
+
+	resp := s.reqAs(t, instanceKey, http.MethodPost, "/api/service-accounts",
+		map[string]any{"name": "team-a-sa", "scope": models.ServiceAccountScopeTeam, "team_id": teamA})
+	if resp.StatusCode != http.StatusCreated {
+		t.Fatalf("create team-scoped account: %d", resp.StatusCode)
+	}
+	var created struct {
+		ServiceAccount struct {
+			ID int64 `json:"id"`
+		} `json:"service_account"`
+	}
+	decode(t, resp, &created)
+
+	// Simulates the adopt-on-409 recovery path: this instance-scoped caller
+	// did not just create this account in this call (a fresh *tdclient.Client
+	// request, same as a second, independent reconcile would issue), yet
+	// still needs to mint it a fresh key.
+	rotateResp := s.reqAs(t, instanceKey, http.MethodPost,
+		"/api/service-accounts/"+id64(created.ServiceAccount.ID)+"/keys", map[string]string{"name": "adopted"})
+	if rotateResp.StatusCode != http.StatusCreated {
+		t.Errorf("instance-scoped account adopting a team-scoped account's key: %d", rotateResp.StatusCode)
+	}
+	rotateResp.Body.Close()
+}
+
+// ---------------------------------------------------------------------------
+// AdminOnly / requireSelfOrAdmin — unchanged after the Caller refactor
+// ---------------------------------------------------------------------------
+
+// The Caller abstraction must not have widened AdminOnly/requireSelfOrAdmin:
+// user management and /api/admin/settings stay human-only, for every scope
+// of service account, exactly as before.
+func TestAdminOnly_RefusesEveryServiceAccountScope(t *testing.T) {
+	s := newTS(t)
+	instanceKey := createServiceAccount(t, s, s.key, "terdut-operator", models.ServiceAccountScopeInstance, 0)
+	teamA := createTeamAs(t, s, instanceKey, "team-a")
+	teamKey := createServiceAccount(t, s, instanceKey, "team-a-sa", models.ServiceAccountScopeTeam, teamA)
+
+	for _, key := range []string{instanceKey, teamKey} {
+		if resp := s.reqAs(t, key, http.MethodGet, "/api/admin/settings", nil); resp.StatusCode != http.StatusForbidden {
+			t.Errorf("expected 403 for a service account reading /api/admin/settings, got %d", resp.StatusCode)
+		} else {
+			resp.Body.Close()
+		}
+		if resp := s.reqAs(t, key, http.MethodPost, "/api/users",
+			map[string]string{"username": "nope", "email": "nope@example.com"}); resp.StatusCode != http.StatusForbidden {
+			t.Errorf("expected 403 for a service account creating a user, got %d", resp.StatusCode)
+		} else {
+			resp.Body.Close()
+		}
+		if resp := s.reqAs(t, key, http.MethodGet, "/api/me", nil); resp.StatusCode != http.StatusForbidden {
+			t.Errorf("expected 403 for a service account calling /api/me, got %d", resp.StatusCode)
+		} else {
+			resp.Body.Close()
+		}
+	}
+}
+
 func TestServiceAccount_SelfRotatesItsOwnKey(t *testing.T) {
 	s := newTS(t)
 	instanceKey := createServiceAccount(t, s, s.key, "terdut-operator", models.ServiceAccountScopeInstance, 0)

internal/api/signup.go

@@ -359,7 +359,19 @@ func handleCreateInvite(db *sql.DB, publicURL string) http.HandlerFunc {
 			respond(w, http.StatusInternalServerError, errResp("internal error"))
 			return
 		}
-		caller, _ := userFromContext(r.Context())
+		// created_by is nullable (ON DELETE SET NULL) for exactly this
+		// reason: the caller minting an invite is not always a human with a
+		// real users row. A team-scoped service account is owner-equivalent
+		// here (requireTeamOwner above already let it through), and this
+		// must leave created_by NULL for one the same way
+		// handleCreateServiceAccount already does for the analogous case —
+		// an unchecked zero value would violate the users(id) foreign key
+		// instead of recording "nobody" cleanly.
+		var createdBy *int64
+		if u, ok := userFromContext(r.Context()); ok {
+			id := u.ID
+			createdBy = &id
+		}
 		expires := time.Now().Add(inviteTTL)
 
 		var out inviteJSON
@@ -368,7 +380,7 @@ func handleCreateInvite(db *sql.DB, publicURL string) http.HandlerFunc {
 			INSERT INTO invites (token_hash, team_id, role, created_by, expires_at, max_uses)
 			VALUES ($1, $2, $3, $4, $5, $6)
 			RETURNING id, team_id, role, created_at, expires_at, max_uses, uses`,
-			hash, teamID, req.Role, caller.ID, expires.Unix(), req.MaxUses).
+			hash, teamID, req.Role, createdBy, expires.Unix(), req.MaxUses).
 			Scan(&out.ID, &out.TeamID, &out.Role, &created, &expiresAt, &out.MaxUses, &out.Uses); err != nil {
 			respond(w, http.StatusInternalServerError, errResp("internal error"))
 			return
@@ -429,7 +441,11 @@ func handleTestNotification(cfg NotifyConfig, db *sql.DB) http.HandlerFunc {
 				errResp("this server has no ntfy configured, so it can send nothing"))
 			return
 		}
-		caller, _ := userFromContext(r.Context())
+		caller, ok := userFromContext(r.Context())
+		if !ok {
+			respond(w, http.StatusForbidden, errResp("this endpoint is for human accounts only"))
+			return
+		}
 
 		var topic *string
 		if err := db.QueryRowContext(r.Context(),
@@ -470,7 +486,11 @@ func handleDismissOnboarding(db *sql.DB) http.HandlerFunc {
 			respond(w, http.StatusBadRequest, errResp("dismissed is required"))
 			return
 		}
-		caller, _ := userFromContext(r.Context())
+		caller, ok := userFromContext(r.Context())
+		if !ok {
+			respond(w, http.StatusForbidden, errResp("this endpoint is for human accounts only"))
+			return
+		}
 
 		var err error
 		if *req.Dismissed {

internal/api/signup_test.go

@@ -2,10 +2,14 @@ package api_test
 
 import (
 	"bytes"
+	"database/sql"
 	"encoding/json"
 	"net/http"
 	"net/http/cookiejar"
 	"testing"
+
+	"git.ryuvia.com/niklas/terdut-server/internal/api"
+	"git.ryuvia.com/niklas/terdut-server/internal/models"
 )
 
 // signup posts to the unauthenticated sign-up endpoint, the way the form does,
@@ -268,3 +272,55 @@ func TestSignup_ValidatesLikeTheRestOfTheServer(t *testing.T) {
 		t.Errorf("an existing username: expected 409, got %d", taken.StatusCode)
 	}
 }
+
+// A team-scoped service account has no users row to attribute created_by to.
+// Before this fix, handleCreateInvite wrote its zero-value caller.ID straight
+// into that (nullable, ON DELETE SET NULL) foreign key instead of leaving it
+// NULL the way handleCreateServiceAccount already does for the same
+// situation — a 500, not the 201 TestServiceAccount_TeamScopeManagesItsOwnInvites
+// now confirms. This pins the column itself ends up NULL, not just "some
+// response came back".
+func TestSignup_InviteCreatedByAServiceAccountLeavesCreatedByNull(t *testing.T) {
+	s := newTS(t)
+	instanceKey := createServiceAccount(t, s, s.key, "terdut-operator", models.ServiceAccountScopeInstance, 0)
+	teamA := createTeamAs(t, s, instanceKey, "team-a")
+	keyA := createServiceAccount(t, s, instanceKey, "team-a-sa", models.ServiceAccountScopeTeam, teamA)
+
+	var created struct {
+		ID int64 `json:"id"`
+	}
+	decode(t, s.reqAs(t, keyA, http.MethodPost, "/api/teams/"+id64(teamA)+"/invites", map[string]any{}), &created)
+
+	var createdBy sql.NullInt64
+	if err := s.db.QueryRow("SELECT created_by FROM invites WHERE id = $1", created.ID).Scan(&createdBy); err != nil {
+		t.Fatalf("read back invites.created_by: %v", err)
+	}
+	if createdBy.Valid {
+		t.Errorf("expected created_by to be NULL for a service-account-minted invite, got %d", createdBy.Int64)
+	}
+}
+
+// Neither of these has a real user_id to act on behalf of; both must 403 a
+// service account explicitly rather than 500 (handleTestNotification, which
+// used to query ntfy_topic for user id 0) or silently no-op (handleDismissOnboarding,
+// which used to UPDATE ... WHERE id = 0, affecting nothing and still
+// returning 204).
+func TestServiceAccount_HumanOnlyEndpointsRefuseExplicitly(t *testing.T) {
+	// BaseURL set (even to a fake, unreachable address) so handleTestNotification
+	// reaches its AsHuman() check instead of short-circuiting on "ntfy not
+	// configured" first — this test is about the human-only check, not ntfy.
+	s := newTS(t, api.NotifyConfig{BaseURL: "http://ntfy.invalid"})
+	instanceKey := createServiceAccount(t, s, s.key, "terdut-operator", models.ServiceAccountScopeInstance, 0)
+
+	if resp := s.reqAs(t, instanceKey, http.MethodPost, "/api/me/notify/test", nil); resp.StatusCode != http.StatusForbidden {
+		t.Errorf("expected 403 for a service account testing notifications, got %d", resp.StatusCode)
+	} else {
+		resp.Body.Close()
+	}
+	if resp := s.reqAs(t, instanceKey, http.MethodPut, "/api/me/onboarding",
+		map[string]bool{"dismissed": true}); resp.StatusCode != http.StatusForbidden {
+		t.Errorf("expected 403 for a service account dismissing onboarding, got %d", resp.StatusCode)
+	} else {
+		resp.Body.Close()
+	}
+}

internal/api/teams.go

@@ -13,12 +13,19 @@ import (
 	"github.com/go-chi/chi/v5"
 )
 
-// handleListTeams lists the caller's own teams, each with their role in it. An
-// administrator listing every team goes through the admin endpoint instead:
-// this one answers "what am I part of", which is what the UI's team filter and
-// the combined queue are built from.
+// handleListTeams lists the caller's own teams, each with their role in it,
+// or — with ?name= — looks up one team by exact name regardless of caller
+// identity (TEAM-LOOKUP.md). An administrator listing every team goes
+// through the admin endpoint instead: the no-name case here answers "what am
+// I part of", which is what the UI's team filter and the combined queue are
+// built from.
 func handleListTeams(db *sql.DB) http.HandlerFunc {
 	return func(w http.ResponseWriter, r *http.Request) {
+		if name := strings.TrimSpace(r.URL.Query().Get("name")); name != "" {
+			handleListTeamsByName(db, w, r, name)
+			return
+		}
+
 		caller, _ := userFromContext(r.Context())
 		rows, err := db.QueryContext(r.Context(), `
 			SELECT t.id, t.name, t.created_at, m.role, m.source
@@ -51,6 +58,31 @@ func handleListTeams(db *sql.DB) http.HandlerFunc {
 	}
 }
 
+// handleListTeamsByName answers "is there a team named exactly this", open to
+// any authenticated caller including a service account (TEAM-LOOKUP.md) —
+// mirrors handleListServiceAccounts' own ?name= lookup: a one-or-zero-length
+// array, never an error on no match, and no caller-identity filtering at
+// all, since what it discloses (a name is taken, nothing about who's in it
+// or any of its data) is the same low sensitivity that lookup already
+// accepts for service-account names.
+func handleListTeamsByName(db *sql.DB, w http.ResponseWriter, r *http.Request, name string) {
+	var t models.Team
+	var created int64
+	err := db.QueryRowContext(r.Context(),
+		"SELECT id, name, created_at FROM teams WHERE name = $1", name,
+	).Scan(&t.ID, &t.Name, &created)
+	if errors.Is(err, sql.ErrNoRows) {
+		respond(w, http.StatusOK, []models.Team{})
+		return
+	}
+	if err != nil {
+		respond(w, http.StatusInternalServerError, errResp("internal error"))
+		return
+	}
+	t.CreatedAt = time.Unix(created, 0).UTC()
+	respond(w, http.StatusOK, []models.Team{t})
+}
+
 // handleUserTeams lists one user's teams, for the admin page's per-user view:
 // "what is this person in", which /api/teams cannot answer because it is always
 // about the caller.

internal/api/teams_test.go

@@ -6,6 +6,8 @@ import (
 	"io"
 	"net/http"
 	"testing"
+
+	"git.ryuvia.com/niklas/terdut-server/internal/models"
 )
 
 // The whole point of #4: two teams sharing one server must not see each other's
@@ -454,3 +456,75 @@ func TestTeams_OutsiderSeesNothing(t *testing.T) {
 		t.Errorf("blue's team list: %v", teams)
 	}
 }
+
+// ---------------------------------------------------------------------------
+// GET /api/teams?name= (TEAM-LOOKUP.md)
+// ---------------------------------------------------------------------------
+
+func TestListTeamsByName_FindsExactMatch(t *testing.T) {
+	s := newTS(t)
+	instanceKey := createServiceAccount(t, s, s.key, "terdut-operator", models.ServiceAccountScopeInstance, 0)
+	teamID := createTeamAs(t, s, instanceKey, "platform")
+
+	teams := list(t, s.reqAs(t, instanceKey, http.MethodGet, "/api/teams?name=platform", nil))
+	if len(teams) != 1 {
+		t.Fatalf("expected exactly one match for ?name=platform, got %d: %v", len(teams), teams)
+	}
+	if int64(teams[0]["id"].(float64)) != teamID {
+		t.Errorf("id = %v, want %d", teams[0]["id"], teamID)
+	}
+	// No membership, so no role to report (models.Team's own doc comment:
+	// "empty when nobody in particular is asking").
+	if _, has := teams[0]["role"]; has {
+		t.Errorf("expected no role on a name-lookup match, got %v", teams[0]["role"])
+	}
+}
+
+func TestListTeamsByName_NoMatchIsAnEmptyArrayNotAnError(t *testing.T) {
+	s := newTS(t)
+	instanceKey := createServiceAccount(t, s, s.key, "terdut-operator", models.ServiceAccountScopeInstance, 0)
+
+	resp := s.reqAs(t, instanceKey, http.MethodGet, "/api/teams?name=does-not-exist", nil)
+	if resp.StatusCode != http.StatusOK {
+		t.Fatalf("expected 200 on no match, got %d", resp.StatusCode)
+	}
+	teams := list(t, resp)
+	if len(teams) != 0 {
+		t.Errorf("expected an empty array, got %v", teams)
+	}
+}
+
+// The actual motivating scenario (TEAM-LOOKUP.md): a service account that
+// already created a team, interrupted before it could remember the id,
+// recovers it via ?name= on the same name its own POST 409s on.
+func TestListTeamsByName_RecoversAfterCreateConflict(t *testing.T) {
+	s := newTS(t)
+	instanceKey := createServiceAccount(t, s, s.key, "terdut-operator", models.ServiceAccountScopeInstance, 0)
+	original := createTeamAs(t, s, instanceKey, "recovered")
+
+	conflict := s.reqAs(t, instanceKey, http.MethodPost, "/api/teams", map[string]string{"name": "recovered"})
+	if conflict.StatusCode != http.StatusConflict {
+		t.Fatalf("expected 409 recreating the same name, got %d", conflict.StatusCode)
+	}
+	conflict.Body.Close()
+
+	teams := list(t, s.reqAs(t, instanceKey, http.MethodGet, "/api/teams?name=recovered", nil))
+	if len(teams) != 1 || int64(teams[0]["id"].(float64)) != original {
+		t.Fatalf("expected to recover the original team %d via ?name=, got %v", original, teams)
+	}
+}
+
+// Not gated by isInstanceServiceAccount or AdminOnly (TEAM-LOOKUP.md): any
+// authenticated caller may ask whether a name is taken, the same low
+// sensitivity GET /api/service-accounts?name= already accepts.
+func TestListTeamsByName_OpenToAnyAuthenticatedCaller(t *testing.T) {
+	s := newTS(t)
+	red := newTeam(t, s, "red")
+	_ = createTeamAs(t, s, s.key, "blue-target")
+
+	// red's own member, not a member of "blue-target", still gets a match.
+	teams := list(t, red.call(http.MethodGet, "/api/teams?name=blue-target", nil))
+	if len(teams) != 1 || teams[0]["name"] != "blue-target" {
+		t.Errorf("expected a non-member caller to still find the team by name, got %v", teams)
+	}
+}

internal/db/db.go

@@ -1,6 +1,7 @@
 package db
 
 import (
+	"context"
 	"database/sql"
 	"embed"
 	"fmt"
@@ -62,6 +63,16 @@ func Open(dsn string) (*sql.DB, error) {
 	}
 }
 
+// migrationLockKey is the Postgres advisory lock Migrate holds for its whole
+// run. Two replicas starting at once would otherwise race the check-then-apply
+// loop below against schema_migrations: the loser could crash on a
+// duplicate-key insert, or contend with the winner's uncommitted DDL. Blocking
+// (pg_advisory_lock, not pg_try_advisory_lock as the archiver and notifier
+// use): on boot there is no later tick to defer to, so the right behaviour is
+// to wait for the other replica to finish migrating, not to skip ahead and
+// start serving against an unmigrated schema.
+const migrationLockKey int64 = 7265_0003
+
 // Migrate applies every embedded migration that has not been applied yet, in
 // filename order, recording each in schema_migrations.
 //
@@ -69,6 +80,22 @@ func Open(dsn string) (*sql.DB, error) {
 // migration that failed half way used to leave the schema in whatever state it
 // had reached. Postgres has transactional DDL, so the rollback is real.
 func Migrate(db *sql.DB) error {
+	ctx := context.Background()
+	conn, err := db.Conn(ctx)
+	if err != nil {
+		return fmt.Errorf("migrate: acquire connection: %w", err)
+	}
+	defer conn.Close()
+
+	if _, err := conn.ExecContext(ctx, "SELECT pg_advisory_lock($1)", migrationLockKey); err != nil {
+		return fmt.Errorf("migrate: acquire advisory lock: %w", err)
+	}
+	defer func() {
+		if _, err := conn.ExecContext(ctx, "SELECT pg_advisory_unlock($1)", migrationLockKey); err != nil {
+			log.Printf("migrate: release advisory lock: %v", err)
+		}
+	}()
+
 	if _, err := db.Exec(`CREATE TABLE IF NOT EXISTS schema_migrations (
 		version    TEXT PRIMARY KEY,
 		applied_at BIGINT NOT NULL DEFAULT FLOOR(EXTRACT(EPOCH FROM now()))::bigint

internal/db/db_test.go

@@ -0,0 +1,125 @@
+package db_test
+
+import (
+	"database/sql"
+	"fmt"
+	"net/url"
+	"os"
+	"strings"
+	"sync"
+	"testing"
+
+	"git.ryuvia.com/niklas/terdut-server/internal/db"
+
+	_ "github.com/jackc/pgx/v5/stdlib"
+)
+
+// TERDUT_TEST_DSN must point at a database the test role may create schemas
+// in; see internal/api/testdb_test.go for the fuller rationale this mirrors.
+// An unset DSN fails rather than skips, deliberately.
+const testDSNEnv = "TERDUT_TEST_DSN"
+
+// TestMigrate_ConcurrentCallersDoNotRace reproduces two replicas starting at
+// once against a brand-new, unmigrated schema: both call db.Migrate at the
+// same time. Before migrationLockKey, the loser could crash on a
+// duplicate-key insert into schema_migrations, or contend with the winner's
+// uncommitted DDL; with the advisory lock, one blocks until the other
+// finishes and both return cleanly.
+func TestMigrate_ConcurrentCallersDoNotRace(t *testing.T) {
+	dsn := os.Getenv(testDSNEnv)
+	if dsn == "" {
+		t.Fatalf("%s is not set: these tests need Postgres.\n"+
+			"Run `make test-db` for a local one, then\n"+
+			"  export %s=postgres://terdut:terdut@localhost:5432/terdut_test?sslmode=disable",
+			testDSNEnv, testDSNEnv)
+	}
+
+	schema := fmt.Sprintf("migrate_race_%d", os.Getpid())
+	admin, err := sql.Open("pgx", dsn)
+	if err != nil {
+		t.Fatalf("connect to %s: %v", testDSNEnv, err)
+	}
+	defer admin.Close()
+	if _, err := admin.Exec("CREATE SCHEMA " + schema); err != nil {
+		t.Fatalf("create schema %s: %v", schema, err)
+	}
+	t.Cleanup(func() {
+		cleanup, err := sql.Open("pgx", dsn)
+		if err != nil {
+			return
+		}
+		defer cleanup.Close()
+		if _, err := cleanup.Exec("DROP SCHEMA " + schema + " CASCADE"); err != nil {
+			t.Logf("drop schema %s: %v", schema, err)
+		}
+	})
+
+	scoped := withSearchPath(dsn, schema)
+
+	const callers = 2
+	errs := make([]error, callers)
+	var wg sync.WaitGroup
+	for i := range callers {
+		wg.Add(1)
+		go func(i int) {
+			defer wg.Done()
+			database, err := db.Open(scoped)
+			if err != nil {
+				errs[i] = fmt.Errorf("open: %w", err)
+				return
+			}
+			defer database.Close()
+			errs[i] = db.Migrate(database)
+		}(i)
+	}
+	wg.Wait()
+
+	for i, err := range errs {
+		if err != nil {
+			t.Fatalf("Migrate #%d: %v", i, err)
+		}
+	}
+
+	entries, err := os.ReadDir("migrations")
+	if err != nil {
+		t.Fatalf("read migrations dir: %v", err)
+	}
+	var want int
+	for _, e := range entries {
+		if !e.IsDir() && strings.HasSuffix(e.Name(), ".sql") {
+			want++
+		}
+	}
+
+	check, err := sql.Open("pgx", scoped)
+	if err != nil {
+		t.Fatalf("connect for verification: %v", err)
+	}
+	defer check.Close()
+
+	var got int
+	if err := check.QueryRow("SELECT COUNT(*) FROM schema_migrations").Scan(&got); err != nil {
+		t.Fatalf("count schema_migrations: %v", err)
+	}
+	if got != want {
+		t.Fatalf("schema_migrations has %d row(s) after two concurrent Migrate calls, want %d (one per migration file, no duplicates)", got, want)
+	}
+}
+
+// withSearchPath pins a DSN to one schema. Copied from
+// internal/api/testdb_test.go rather than shared: that helper lives in the
+// api_test package, a separate compiled package this one cannot import.
+func withSearchPath(dsn, schema string) string {
+	opt := "-csearch_path=" + schema
+
+	if strings.HasPrefix(dsn, "postgres://") || strings.HasPrefix(dsn, "postgresql://") {
+		u, err := url.Parse(dsn)
+		if err == nil {
+			q := u.Query()
+			q.Set("options", opt)
+			u.RawQuery = q.Encode()
+			return u.String()
+		}
+	}
+	return dsn + " options='" + opt + "'"
+}

internal/web/static/app.css

@@ -47,12 +47,27 @@
   --font: system-ui, -apple-system, "Segoe UI", Roboto, "Helvetica Neue", Arial, sans-serif;
   --mono: ui-monospace, SFMono-Regular, Menlo, Consolas, monospace;
 
+  /* A ~4-size type scale, per issue #26's "use one type scale" ask. This
+     file still has a dozen one-off font-size values below; migrating the
+     low-risk, purely cosmetic ones (standalone titles with no dimensional
+     or functional constraint) onto these tokens is a start, not the whole
+     job — the rest (sizes tied to a fixed shape like the avatar circle, to
+     a deliberately prominent display like a stat tile or the on-call name,
+     or to a non-negotiable constraint like the 16px that stops iOS zooming
+     into an input) stay as either their own pixel value or a documented
+     exception, since guessing at those without seeing them render risks
+     trading one inconsistency for a worse one. */
+  --fs-xs: 12px;
+  --fs-sm: 13px;
+  --fs-base: 14px;
+  --fs-lg: 18px;
+  --fs-xl: 21px;
+
   --topbar-h: 52px;
-  /* No bottom tab bar on any breakpoint any more — mobile uses the hamburger
-     menu in the topbar, desktop the sidebar — so this stays 0. Kept as a
-     variable rather than deleted since .view, .toast and .nav's own height
-     calc still read it. */
-  --tabbar-h: 0px;
+  /* The phone-width bottom tab bar's height. Unused above 900px: the
+     desktop block overrides .nav/.view/.toast directly rather than reading
+     this back down to 0. */
+  --tabbar-h: 58px;
   --safe-top: env(safe-area-inset-top, 0px);
   --safe-bottom: env(safe-area-inset-bottom, 0px);
 }
@@ -212,9 +227,6 @@ input:focus, textarea:focus { outline: none; border-color: var(--accent); box-sh
   font-size: 18px; font-weight: 700; letter-spacing: -0.01em;
   overflow: hidden; text-overflow: ellipsis; white-space: nowrap; min-width: 0;
 }
-#menu-btn { position: relative; }
-.nav-badge.menu-btn-badge { top: 2px; left: auto; right: 2px; }
-
 .open-pill {
   display: inline-flex; align-items: center; gap: 6px;
   padding: 3px 10px; border-radius: 999px;
@@ -226,11 +238,11 @@ input:focus, textarea:focus { outline: none; border-color: var(--accent); box-sh
 .open-pill.has-triggered::before { background: var(--crit); }
 .open-pill.all-acked::before { background: var(--warn); }
 
-/* Hidden on phones — mobile navigates through the hamburger menu in the
-   topbar instead (see #menu-btn / openNavMenu in app.js). Reappears as the
-   left sidebar from 900px, where the desktop block below redeclares display. */
+/* Bottom tab bar on phones (Queue, On-call, Alerts, Team, More); becomes the
+   left sidebar from 900px, where the desktop block below redeclares display
+   and shows every section flat, .nav-link-secondary included. */
 .nav {
-  display: none;
+  display: flex;
   position: fixed; left: 0; right: 0; bottom: 0; z-index: 20;
   height: calc(var(--tabbar-h) + var(--safe-bottom));
   padding-bottom: var(--safe-bottom);
@@ -240,14 +252,24 @@ input:focus, textarea:focus { outline: none; border-color: var(--accent); box-sh
   border-top: 1px solid var(--border);
 }
 .nav-brand { display: none; }
+/* Hidden here (shown from 900px below): on the phone bar the team switcher
+   lives in the topbar instead, as #team-selector-mobile. */
+.nav-team-selector { display: none; }
 .nav-link {
   position: relative;
+  /* flex: 1 spreads the tabs evenly across the bar's width; the desktop
+     block below cancels it back to a natural-width row item. */
+  flex: 1 1 0;
   display: flex; flex-direction: column; align-items: center; justify-content: center; gap: 2px;
   color: var(--faint); font-size: 11px; font-weight: 600;
   /* min-width lets a column shrink below its label's natural width, which is
      what stops six tabs widening the bar past the screen. */
   min-width: 0; padding: 0 2px;
 }
+/* Must come after .nav-link above: same specificity (one class each), so
+   whichever is later in the file wins for an element wearing both classes,
+   and this needs to beat .nav-link's display:flex here on the phone bar. */
+.nav-link-secondary { display: none; }
 .nav-label {
   max-width: 100%; overflow: hidden; text-overflow: ellipsis; white-space: nowrap;
 }
@@ -276,6 +298,7 @@ input:focus, textarea:focus { outline: none; border-color: var(--accent); box-sh
 }
 .team-selector-mobile { padding: 4px 10px; font-size: 12px; max-width: 120px; }
 .team-selector-label { overflow: hidden; text-overflow: ellipsis; white-space: nowrap; }
+.team-selector-chevron { width: 14px; height: 14px; flex: none; color: var(--faint); margin-left: -2px; }
 
 /* A team's identity colour — not a status, so never the severity palette. Six
    colours, then they repeat; teamColorClass() in format.js picks one by the
@@ -303,6 +326,7 @@ input:focus, textarea:focus { outline: none; border-color: var(--accent); box-sh
 /* ---------- chips ---------- */
 
 .chips {
+  position: relative;
   display: flex; gap: 6px;
   padding: 12px 16px 8px;
   overflow-x: auto; scrollbar-width: none;
@@ -318,6 +342,17 @@ input:focus, textarea:focus { outline: none; border-color: var(--accent); box-sh
 }
 .chip[aria-selected="true"] { background: var(--text); border-color: var(--text); color: var(--bg); }
 .chip .count { margin-left: 4px; opacity: 0.7; }
+/* An overlay, not a flex item: absolute against .chips' own (non-scrolling)
+   box stays flush with its real right edge regardless of scroll position,
+   which turned out not to be true of position:sticky here — as a flex
+   item, its sticky offset interacted with the row's gap and its own
+   negative margin, landing short of the edge by about one gap's width. */
+.chips-fade {
+  position: absolute; top: 0; right: 0; bottom: 0;
+  width: 24px;
+  background: linear-gradient(to right, transparent, var(--bg));
+  pointer-events: none;
+}
 
 /* ---------- lists ---------- */
 
@@ -381,7 +416,7 @@ input:focus, textarea:focus { outline: none; border-color: var(--accent); box-sh
 .empty {
   padding: 48px 16px; text-align: center; color: var(--muted);
 }
-.empty strong { display: block; color: var(--text); font-size: 16px; margin-bottom: 4px; }
+.empty strong { display: block; color: var(--text); font-size: var(--fs-lg); margin-bottom: 4px; }
 .empty .icon { width: 36px; height: 36px; color: var(--ok); margin-bottom: 8px; }
 
 .load-error {
@@ -434,8 +469,12 @@ input:focus, textarea:focus { outline: none; border-color: var(--accent); box-sh
 .detail-head .copy { margin-left: auto; }
 .clip-buffer { position: fixed; top: 0; left: 0; opacity: 0; pointer-events: none; }
 .detail-head .crumb { font-weight: 600; color: var(--muted); font-size: 14px; }
-.detail-title { font-size: 21px; font-weight: 750; letter-spacing: -0.01em; margin: 16px 0 8px; overflow-wrap: anywhere; }
+.detail-title { font-size: var(--fs-xl); font-weight: 750; letter-spacing: -0.01em; margin: 16px 0 8px; overflow-wrap: anywhere; }
 .detail-badges { display: flex; flex-wrap: wrap; gap: 6px; margin-bottom: 14px; }
+/* A copy of the sticky actionbar's primary button, right under the status
+   it responds to — see quickActions() in incident.js. */
+.detail-quick-actions { margin-bottom: 14px; }
+.detail-quick-actions .btn-primary { font-size: 16px; min-height: 44px; }
 
 .card {
   background: var(--surface);
@@ -450,6 +489,9 @@ input:focus, textarea:focus { outline: none; border-color: var(--accent); box-sh
 .facts dt { color: var(--muted); }
 .facts dd { margin: 0; overflow-wrap: anywhere; }
 .facts .sub { color: var(--faint); }
+.facts dd.fact-summary { display: flex; flex-wrap: wrap; align-items: center; gap: 8px; }
+.fact-chip { display: inline-flex; align-items: center; gap: 4px; }
+.fact-icon { width: 15px; height: 15px; color: var(--faint); }
 
 .section { margin-top: 22px; }
 .section-title {
@@ -481,6 +523,17 @@ details > summary::before { content: "▸ "; }
 details[open] > summary::before { content: "▾ "; }
 details[open] > summary { margin-bottom: 8px; }
 
+/* One .tl-phase per status the incident has been through (see
+   timelinePhases() in incident.js) — each with its own .timeline <ol>, so
+   the existing :first-child/:last-child rail-capping below gives each phase
+   its own self-contained connecting line rather than one running through
+   the headings. */
+.tl-phase + .tl-phase { border-top: 1px solid var(--border); }
+.tl-phase-title {
+  padding: 10px 14px 0;
+  font-size: 11px; font-weight: 700; text-transform: uppercase; letter-spacing: 0.06em;
+  color: var(--faint);
+}
 .timeline { list-style: none; margin: 0; padding: 4px 0; }
 .tl-item {
   position: relative;
@@ -513,7 +566,7 @@ details[open] > summary { margin-bottom: 8px; }
 }
 .note-fix { background: var(--ok-soft); border-left: 3px solid var(--ok); }
 .similar { list-style: none; margin: 0; padding: 0; }
-.similar-item { padding: 10px 0; }
+.similar-item { padding: 10px 14px; }
 .similar-item + .similar-item { border-top: 1px solid var(--border, var(--surface-2)); }
 .check { display: flex; align-items: center; gap: 8px; font-size: 14px; color: var(--muted); }
 .note-actions { display: flex; justify-content: flex-end; }
@@ -552,7 +605,7 @@ details[open] > summary { margin-bottom: 8px; }
 @keyframes sheet-up { from { transform: translateY(24px); opacity: 0.6; } }
 .sheet-inner { padding: 8px 16px calc(16px + var(--safe-bottom)); }
 .sheet-grab { width: 40px; height: 4px; margin: 0 auto 12px; border-radius: 2px; background: var(--border-strong); }
-.sheet-title { font-size: 17px; font-weight: 700; margin: 0 0 4px; }
+.sheet-title { font-size: var(--fs-lg); font-weight: 700; margin: 0 0 4px; }
 .sheet-text { color: var(--muted); margin: 0 0 14px; font-size: 14px; }
 .sheet-form { display: grid; gap: 12px; }
 .sheet-actions { display: flex; gap: 8px; margin-top: 16px; }
@@ -604,9 +657,18 @@ details[open] > summary { margin-bottom: 8px; }
 .now-label { color: var(--muted); font-size: 13px; font-weight: 600; }
 .now-name { font-size: 20px; font-weight: 750; }
 .you { color: var(--accent); font-weight: 650; font-size: 13px; margin-left: 6px; }
+/* Oncall's own "you" indicator only (see you() in oncall.js) — a pill badge
+   is easier to spot there than this plain accent-coloured text. */
+.you-badge { margin-left: 6px; }
 
 .week-nav { display: flex; align-items: center; gap: 4px; }
 .week-nav .label { font-size: 14px; font-weight: 650; min-width: 9em; text-align: center; }
+/* The week-nav button that shows the date range, "28 Sep – 4 Oct", with the
+   ISO week number as secondary text inside it — a separate class from
+   .label above (team.js's month-nav uses that one) so its <small> isn't
+   caught by the unrelated .label > span styling meant for label chips. */
+.week-nav .week-label { font-size: 14px; font-weight: 650; min-width: 11.5em; text-align: center; white-space: nowrap; }
+.week-label small { color: var(--faint); font-weight: 600; font-size: 11px; margin-left: 2px; }
 .days { list-style: none; margin: 0; padding: 0; }
 .day { display: grid; grid-template-columns: 3.2em 4.2em 1fr; align-items: center; gap: 8px; min-height: 50px; padding: 0 14px; }
 .day + .day { border-top: 1px solid var(--border); }
@@ -614,11 +676,18 @@ details[open] > summary { margin-bottom: 8px; }
 .day-date { color: var(--faint); font-size: 13px; }
 .day-who { overflow: hidden; text-overflow: ellipsis; white-space: nowrap; }
 .day-who.nobody { color: var(--faint); font-style: italic; }
+/* A run of several days held by the same person (or left empty), replacing
+   what used to be one identical row per day — see weekRuns() in oncall.js. */
+.day.range { grid-template-columns: 1fr auto; }
+.day-range { font-weight: 650; }
 .day.today { background: var(--accent-soft); }
 .day.today:first-child { border-radius: var(--radius) var(--radius) 0 0; }
 .day.today:last-child { border-radius: 0 0 var(--radius) var(--radius); }
-.day.today .day-name { color: var(--accent); }
+.day.today .day-name, .day.today .day-range { color: var(--accent); }
 .day.past { opacity: 0.6; }
+/* Highlights whichever row is yours, same soft tint as .today — they already
+   read fine layered (today's own row is almost always one of yours too). */
+.day.mine { background: var(--accent-soft); }
 
 .shift-list { list-style: none; margin: 0; padding: 0; }
 .shift-list li { display: flex; justify-content: space-between; padding: 12px 14px; }
@@ -675,9 +744,11 @@ kbd {
     display: flex; align-items: center; gap: 10px;
     padding: 4px 10px 18px; font-size: 18px; font-weight: 750; letter-spacing: -0.01em;
   }
-  .nav-team-selector { margin: -8px 10px 14px; width: calc(100% - 20px); }
+  .nav-team-selector { display: inline-flex; margin: -8px 10px 14px; width: calc(100% - 20px); }
+  .nav-link-secondary { display: flex; }
+  .nav-more-btn { display: none; }
   .nav-link {
-    flex-direction: row; justify-content: flex-start; gap: 12px;
+    flex: none; flex-direction: row; justify-content: flex-start; gap: 12px;
     min-height: 40px; padding: 0 10px; border-radius: var(--radius-sm);
     color: var(--muted); font-size: 14px;
   }
@@ -701,6 +772,7 @@ kbd {
      is hidden, so the chips wrap here instead: Archived stays reachable. */
   .pane-list .chips { flex-wrap: wrap; overflow-x: visible; }
   .pane-list .chip-sep { display: none; }
+  .chips-fade { display: none; }
   .view-queue:not(.has-detail) .pane-detail { display: block; }
 
   /* On desktop the list stays visible next to the detail. */
@@ -966,6 +1038,7 @@ button.rota-week:hover { background: var(--surface-2); color: var(--text); }
 .overview-head { display: flex; align-items: baseline; gap: 8px; }
 .overview-count { margin-left: auto; color: var(--muted); font-size: 18px; font-weight: 700; }
 .overview-item p { margin: 4px 0 0; }
+.overview-note-icon { width: 13px; height: 13px; vertical-align: -2px; color: var(--ok); }
 
 /* ---------- stats page ---------- */
 

internal/web/static/index.html

@@ -102,7 +102,10 @@
         <svg viewBox="0 0 24 24" aria-hidden="true"><path d="M6 16V11a6 6 0 0 1 12 0v5l1.5 2h-15z"/><path d="M10 20.5a2 2 0 0 0 4 0"/></svg>
         <span class="nav-label">Alerts</span>
       </a>
-      <a class="nav-link" href="/stats" data-section="stats" aria-label="Stats">
+      <!-- Secondary: full-width in the desktop sidebar, folded into the
+           "More" tab's sheet on the phone-width bottom bar instead (see
+           .nav-link-secondary in app.css and openNavMenu in app.js). -->
+      <a class="nav-link nav-link-secondary" href="/stats" data-section="stats" aria-label="Stats">
         <svg viewBox="0 0 24 24" aria-hidden="true"><path d="M4 20h16M7 20v-7M12 20V6M17 20v-10"/></svg>
         <span class="nav-label">Stats</span>
       </a>
@@ -113,22 +116,26 @@
       <!-- Hidden unless the signed-in user is a system administrator; app.js
            unhides it once /api/me says so. The server refuses every admin
            endpoint regardless, so this is a courtesy and not a gate. -->
-      <a class="nav-link" href="/admin" data-section="admin" aria-label="Admin" id="nav-admin" hidden>
+      <a class="nav-link nav-link-secondary" href="/admin" data-section="admin" aria-label="Admin" id="nav-admin" hidden>
         <svg viewBox="0 0 24 24" aria-hidden="true"><path d="M12 3l7 3v6c0 4-3 7-7 9-4-2-7-5-7-9V6z"/></svg>
         <span class="nav-label">Admin</span>
       </a>
-      <a class="nav-link" href="/more" data-section="more" aria-label="Account">
+      <a class="nav-link nav-link-secondary" href="/more" data-section="more" aria-label="Account">
         <svg viewBox="0 0 24 24" aria-hidden="true"><circle cx="12" cy="8" r="3.5"/><path d="M5 20a7 7 0 0 1 14 0"/></svg>
         <span class="nav-label">Account</span>
       </a>
+      <!-- Phone-width only (see .nav-more-btn in app.css): opens the same
+           sheet the old hamburger button did, for the sections the bottom
+           bar has no room for. Not shown on the desktop sidebar, which lists
+           every section already. -->
+      <button class="nav-link nav-more-btn" id="nav-more-btn" type="button" aria-label="More sections" aria-haspopup="menu">
+        <svg viewBox="0 0 24 24" aria-hidden="true"><path d="M5 12h.01M12 12h.01M19 12h.01"/></svg>
+        <span class="nav-label">More</span>
+      </button>
     </nav>
 
     <header class="topbar">
       <div class="topbar-left">
-        <button class="btn btn-ghost btn-icon" id="menu-btn" type="button" aria-label="Menu" aria-haspopup="menu">
-          <svg viewBox="0 0 24 24" aria-hidden="true"><path d="M4 7h16M4 12h16M4 17h16"/></svg>
-          <span class="nav-badge menu-btn-badge" data-badge hidden></span>
-        </button>
         <button class="team-selector-mobile" id="team-selector-mobile" type="button" hidden></button>
         <h1 class="topbar-title" id="topbar-title">Queue</h1>
       </div>

internal/web/static/js/alerts.js

@@ -60,7 +60,11 @@ function render() {
   let body;
   if (error && !items) body = h('div', { class: 'load-error', text: error });
   else if (!items) body = spinner();
-  else if (!items.length) body = emptyState(filter === 'firing' ? 'Nothing firing' : 'No alerts', '', filter === 'firing' ? 'checkCircle' : null);
+  else if (!items.length) {
+    body = filter === 'firing'
+      ? emptyState('Nothing firing', 'No alerts are currently firing.', 'checkCircle')
+      : emptyState('No alerts', 'None match this filter.', 'bell');
+  }
   else {
     body = h('div', { class: 'list' },
       error && h('div', { class: 'load-error', text: `Showing older data: ${error}` }),

internal/web/static/js/app.js

@@ -36,16 +36,17 @@ const SECTIONS = {
   device: { title: 'Sign in a terminal', view: device },
 };
 
-// The mobile hamburger menu's contents — the same sections the desktop
-// sidebar's .nav-link list carries in index.html, in the same order.
+// The desktop sidebar's .nav-link list in index.html, in the same order.
+// `secondary` marks the ones that fold into the phone bottom bar's "More"
+// sheet (openNavMenu below) instead of getting a tab of their own there.
 const NAV_ITEMS = [
   { path: '/', section: 'queue', label: 'Queue', icon: 'queueList' },
   { path: '/oncall', section: 'oncall', label: 'On-call', icon: 'calendar' },
   { path: '/alerts', section: 'alerts', label: 'Alerts', icon: 'bell' },
-  { path: '/stats', section: 'stats', label: 'Stats', icon: 'chart' },
+  { path: '/stats', section: 'stats', label: 'Stats', icon: 'chart', secondary: true },
   { path: '/team', section: 'team', label: 'Team', icon: 'team' },
-  { path: '/admin', section: 'admin', label: 'Admin', icon: 'shield', adminOnly: true },
-  { path: '/more', section: 'more', label: 'Account', icon: 'user' },
+  { path: '/admin', section: 'admin', label: 'Admin', icon: 'shield', adminOnly: true, secondary: true },
+  { path: '/more', section: 'more', label: 'Account', icon: 'user', secondary: true },
 ];
 
 function parseRoute(pathname) {
@@ -152,15 +153,16 @@ function render() {
 
 // ---------- nav menu ----------
 
-// The mobile hamburger menu: same shape as the sheet-based action menus in
-// incident.js (openSheet + a <ul class="menu"> of menu-item buttons), one
-// item per NAV_ITEMS entry, resolving with a path for navigate() to use.
+// The phone bottom bar's "More" sheet: same shape as the sheet-based action
+// menus in incident.js (openSheet + a <ul class="menu"> of menu-item
+// buttons), one item per secondary NAV_ITEMS entry — the ones the bar itself
+// has no room for, since Queue/On-call/Alerts/Team already have their own
+// tab and don't need to be reachable here too.
 function openNavMenu() {
   const current = SECTIONS[route.section].nav || route.section;
-  const triggered = state.open.filter((i) => i.status === 'triggered').length;
-  const items = NAV_ITEMS.filter((n) => !n.adminOnly || state.me?.user?.is_admin);
+  const items = NAV_ITEMS.filter((n) => n.secondary && (!n.adminOnly || state.me?.user?.is_admin));
   ui.openSheet(() => [
-    ui.h('h2', { class: 'sheet-title', text: 'Sections' }),
+    ui.h('h2', { class: 'sheet-title', text: 'More' }),
     ui.h('ul', { class: 'menu', role: 'menu' }, items.map((n) =>
       ui.h('li', {}, ui.h('button', {
         class: 'menu-item',
@@ -171,7 +173,6 @@ function openNavMenu() {
       },
         ui.icon(n.icon),
         n.label,
-        n.section === 'queue' && triggered > 0 && ui.badge(String(triggered), 'st-triggered menu-sub'),
       ))),
     ),
   ]).then((path) => {
@@ -205,8 +206,8 @@ function updateBadges() {
   pill.classList.toggle('has-triggered', triggered > 0);
   pill.classList.toggle('all-acked', open > 0 && triggered === 0);
 
-  // Two badges carry this count: the sidebar's Queue tab (desktop) and the
-  // hamburger button (mobile) — only one of the two is ever visible at once.
+  // Just the Queue tab's own badge now — phone bottom bar and desktop
+  // sidebar both read the same [data-badge] span on that one nav-link.
   for (const badge of document.querySelectorAll('[data-badge]')) {
     badge.hidden = triggered === 0;
     badge.textContent = String(triggered);
@@ -230,7 +231,7 @@ async function boot() {
   document.addEventListener('keydown', onKey);
   $('login-form').addEventListener('submit', onLogin);
   $('signup-form').addEventListener('submit', onSignup);
-  $('menu-btn').addEventListener('click', openNavMenu);
+  $('nav-more-btn').addEventListener('click', openNavMenu);
   teamselector.init();
   ssoErrorCode = takeSSOError();
 
@@ -355,6 +356,7 @@ function ssoErrorText(code, name) {
     no_email: `${sso} did not send an email address for you, which terdut needs.`,
     email_conflict: 'An account with your email address already exists and could not be linked to this sign-in. Ask an administrator.',
     disabled: 'Your account is disabled. Ask an administrator.',
+    not_bootstrapped: 'This install is still setting up. Try again in a moment.',
   }[code] || `Signing in with ${sso} failed.`;
 }
 

internal/web/static/js/incident.js

@@ -7,7 +7,7 @@ import {
   h, clear, icon, badge, labelChip, openSheet, closeSheet, confirm, toast, spinner, emptyState,
 } from './ui.js';
 import {
-  ago, when, until, isFuture, severityClass, STATUS_LABEL,
+  ago, when, until, duration, isFuture, severityClass, STATUS_LABEL,
 } from './format.js';
 import { myID, users } from './state.js';
 import { back } from './app.js';
@@ -86,6 +86,7 @@ function render() {
       error && h('div', { class: 'load-error', text: `Showing older data: ${error}` }),
       h('h1', { class: 'detail-title', text: inc.title }),
       h('div', { class: 'detail-badges' }, statusBadges()),
+      quickActions(),
       facts(),
       groupLabels(),
       alertsSection(),
@@ -125,6 +126,20 @@ function who(id, name) {
 function facts() {
   const rows = [];
   const add = (k, ...v) => rows.push(h('dt', { text: k }), h('dd', {}, ...v));
+
+  // Duration, severity and who's on it, in one scannable row up top — the
+  // rest of this card has each of those too, but spread across rows that
+  // take reading top to bottom to piece together.
+  const elapsedTo = inc.resolved_at ? Date.parse(inc.resolved_at) : Date.now();
+  const responsible = inc.assigned_to_id != null ? who(inc.assigned_to_id, inc.assigned_to)
+    : inc.acknowledged_by_id != null ? who(inc.acknowledged_by_id, inc.acknowledged_by)
+      : 'Unassigned';
+  rows.push(h('dt', { text: 'At a glance' }), h('dd', { class: 'fact-summary' },
+    h('span', { class: 'fact-chip' }, icon('clock', 'icon fact-icon'), duration(elapsedTo - Date.parse(inc.triggered_at))),
+    inc.severity && badge(inc.severity, `plain ${severityClass(inc.severity)}`),
+    h('span', { class: 'fact-chip' }, icon('user', 'icon fact-icon'), responsible),
+  ));
+
   add('Triggered', when(inc.triggered_at), h('span', { class: 'sub', text: ` · ${ago(inc.triggered_at)}` }));
   if (inc.acknowledged_at) {
     add('Acknowledged', `${who(inc.acknowledged_by_id, inc.acknowledged_by)} · ${when(inc.acknowledged_at)}`);
@@ -148,9 +163,12 @@ function facts() {
 function groupLabels() {
   const entries = Object.entries(inc.group_labels || {});
   if (!entries.length) return null;
+  // Collapsed by default, the same disclosure alertItem() below uses for an
+  // alert's own labels — this is background, not something to scan past.
   return h('section', { class: 'section' },
-    h('h2', { class: 'section-title', text: 'Grouped by' }),
-    h('div', { class: 'labels-wrap' }, entries.map(([k, v]) => labelChip(k, v))),
+    h('details', {},
+      h('summary', { text: `Grouped by (${entries.length})` }),
+      h('div', { class: 'labels-wrap' }, entries.map(([k, v]) => labelChip(k, v)))),
   );
 }
 
@@ -206,12 +224,20 @@ function eventText(ev, named = false) {
     case 'snoozed': return [strong(person), ` snoozed until ${ev.detail ? when(ev.detail) : '…'}`];
     case 'unsnoozed': return [strong(person), ' ended the snooze'];
     case 'resolved': return person ? [strong(person), ' resolved the incident'] : ['Resolved: every alert stopped firing'];
+    // Falls through to the generic `${ev.type}: ${ev.detail}` below otherwise
+    // — this just capitalises it and drops the redundant "escalated:" prefix
+    // from detail (already "level 2: alice, bob" or "escalation exhausted: …").
+    case 'escalated': return [`Escalated — ${ev.detail}`];
     case 'note': return [strong(person), ' added a note'];
     case 'resolution_note': return [strong(person), ' noted what fixed it'];
     case 'notified': {
       const to = person ? strong(person) : 'the fallback topic';
       if (ev.detail === 'reminder') return ['Reminder sent to ', to];
       if (ev.detail === 'resolved') return ['Resolution sent to ', to];
+      // 'escalated' is a second, later page — the next level firing, not the
+      // same page landing twice — so it reads as a bug unless told apart
+      // from the initial 'triggered' page below.
+      if (ev.detail === 'escalated') return ['Escalation paged ', to];
       return ['Paged ', to];
     }
     case 'notify_failed': return ['Notification failed', ev.detail ? `: ${ev.detail}` : ''];
@@ -236,8 +262,34 @@ function similarSection() {
   );
 }
 
+// Splits the already-sorted timeline on the status transitions that matter —
+// first acknowledged, then resolved — so a long incident reads as "before
+// anyone had it" / "while someone did" / "after it closed" instead of one
+// undifferentiated list. A later re-acknowledge (after an unacknowledge)
+// doesn't open a second "Acknowledged" phase; it's still the same spell of
+// somebody owning it.
+function timelinePhases(sorted) {
+  const phases = [{ label: 'Triggered', events: [] }];
+  let acked = false;
+  for (const ev of sorted) {
+    if (ev.type === 'acknowledged' && !acked) {
+      phases.push({ label: 'Acknowledged', events: [] });
+      acked = true;
+    } else if (ev.type === 'resolved') {
+      phases.push({ label: 'Resolved', events: [] });
+    }
+    phases[phases.length - 1].events.push(ev);
+  }
+  return phases.filter((p) => p.events.length);
+}
+
 function timelineSection() {
   const sorted = [...events].sort((a, b) => Date.parse(a.created_at) - Date.parse(b.created_at) || a.id - b.id);
+  const phases = timelinePhases(sorted);
+  // A single phase (the common case: most incidents are acked once and
+  // resolved) names nothing extra — only a split timeline needs the
+  // headings to make sense of.
+  const named = phases.length > 1;
   return h('section', { class: 'section' },
     h('h2', { class: 'section-title' },
       h('span', { text: 'Timeline' }),
@@ -245,8 +297,10 @@ function timelineSection() {
         icon('note'), 'Add note')),
     h('div', { class: 'card' },
       sorted.length
-        ? h('ol', { class: 'timeline' }, sorted.map(timelineItem))
-        : emptyState('No events yet', '')),
+        ? phases.map((p) => h('div', { class: 'tl-phase' },
+          named && h('div', { class: 'tl-phase-title', text: p.label }),
+          h('ol', { class: 'timeline' }, p.events.map(timelineItem))))
+        : emptyState('No events yet', 'Nothing has happened on this incident yet.', 'clock')),
   );
 }
 
@@ -365,27 +419,43 @@ async function copyIncident() {
 const isOpen = () => inc.status !== 'resolved';
 const isSnoozed = () => isOpen() && isFuture(inc.snoozed_until);
 
-function actionBar() {
-  let primary;
-  let secondary;
+// primaryAction and secondaryAction are factories, not shared nodes — a
+// button can only live in one place, and quickActions() below needs its own
+// copy of the primary one rather than the actionbar's.
+function primaryAction() {
   if (inc.status === 'triggered') {
-    primary = h('button', { class: 'btn btn-primary', type: 'button', onclick: acknowledge }, icon('check'), 'Acknowledge');
-  } else if (inc.status === 'acknowledged') {
-    primary = h('button', { class: 'btn btn-primary', type: 'button', onclick: resolve }, icon('checkCircle'), 'Resolve');
-  } else {
-    primary = inc.archived_at
-      ? h('button', { class: 'btn btn-primary', type: 'button', onclick: unarchive }, icon('undo'), 'Unarchive')
-      : h('button', { class: 'btn btn-primary', type: 'button', onclick: archive }, icon('archive'), 'Archive');
+    return h('button', { class: 'btn btn-primary', type: 'button', onclick: acknowledge }, icon('check'), 'Acknowledge');
   }
+  if (inc.status === 'acknowledged') {
+    return h('button', { class: 'btn btn-primary', type: 'button', onclick: resolve }, icon('checkCircle'), 'Resolve');
+  }
+  return inc.archived_at
+    ? h('button', { class: 'btn btn-primary', type: 'button', onclick: unarchive }, icon('undo'), 'Unarchive')
+    : h('button', { class: 'btn btn-primary', type: 'button', onclick: archive }, icon('archive'), 'Archive');
+}
+
+function secondaryAction() {
   if (isOpen()) {
-    secondary = isSnoozed()
+    return isSnoozed()
       ? h('button', { class: 'btn', type: 'button', onclick: unsnooze }, icon('bell'), 'Unsnooze')
       : h('button', { class: 'btn', type: 'button', onclick: snooze }, icon('clock'), 'Snooze');
-  } else {
-    secondary = h('button', { class: 'btn', type: 'button', onclick: addNote }, icon('note'), 'Note');
   }
-  const more = h('button', { class: 'btn btn-icon', type: 'button', 'aria-label': 'More actions', onclick: moreMenu }, icon('more'));
-  const bar = h('div', { class: 'actionbar' }, primary, secondary, more);
+  return h('button', { class: 'btn', type: 'button', onclick: addNote }, icon('note'), 'Note');
+}
+
+// A copy of the primary action (Acknowledge/Resolve/…) up where it's seen
+// right away, next to the status it responds to. The sticky actionbar below
+// keeps carrying every action, primary included, for whenever the page has
+// been scrolled past it.
+function quickActions() {
+  const div = h('div', { class: 'detail-quick-actions' }, primaryAction());
+  if (busy) for (const b of div.querySelectorAll('button')) b.disabled = true;
+  return div;
+}
+
+function actionBar() {
+  const more = h('button', { class: 'btn', type: 'button', 'aria-label': 'More actions', onclick: moreMenu }, icon('more'), 'More');
+  const bar = h('div', { class: 'actionbar' }, primaryAction(), secondaryAction(), more);
   if (busy) for (const b of bar.querySelectorAll('button')) b.disabled = true;
   return bar;
 }

internal/web/static/js/oncall.js

@@ -6,8 +6,8 @@
 // every team the viewer is in, because somebody on two rotas wants both.
 
 import * as api from './api.js';
-import { h, clear, icon, spinner } from './ui.js';
-import { isoDate, mondayOf, addDays, isoWeek, initial } from './format.js';
+import { h, clear, badge, icon, spinner } from './ui.js';
+import { isoDate, mondayOf, addDays, isoWeek, initial, duration } from './format.js';
 import { myID, currentTeam } from './state.js';
 
 const view = () => document.getElementById('view-oncall');
@@ -68,7 +68,7 @@ function render() {
 }
 
 function you(userID) {
-  return userID === myID() ? h('span', { class: 'you', text: 'you' }) : null;
+  return userID === myID() ? badge('you', 'plain st-oncall you-badge') : null;
 }
 
 // One card per team with somebody on call, and a single empty card when there
@@ -99,20 +99,48 @@ function nowCard() {
     )));
 }
 
+// Groups the week's 7 days into runs held by the same person (or the same
+// empty slot) — the week's own version of the consecutive-day grouping
+// myShifts does for a single person's own dates, below. Seven identical rows
+// for one person all week collapses to the one bar this way.
+function weekRuns(byDate) {
+  const runs = [];
+  for (let i = 0; i < 7; i++) {
+    const date = isoDate(addDays(weekStart, i));
+    const e = byDate.get(date) || null;
+    const uid = e ? e.user_id : null;
+    const last = runs[runs.length - 1];
+    if (last && last.uid === uid) last.to = date;
+    else runs.push({ uid, entry: e, from: date, to: date });
+  }
+  return runs;
+}
+
 function weekCard() {
   const byDate = new Map(data.week.map((e) => [e.date, e]));
   const today = isoDate(new Date());
-  const days = [];
-  for (let i = 0; i < 7; i++) {
-    const d = addDays(weekStart, i);
-    const key = isoDate(d);
-    const e = byDate.get(key);
-    days.push(h('li', { class: `day ${key === today ? 'today' : ''} ${key < today ? 'past' : ''}` },
-      h('span', { class: 'day-name', text: dayName.format(d) }),
-      h('span', { class: 'day-date', text: dayDate.format(d) }),
-      h('span', { class: `day-who ${e ? '' : 'nobody'}` }, e ? e.username : 'nobody', e && you(e.user_id)),
-    ));
-  }
+  const mine = myID();
+  const days = weekRuns(byDate).map((r) => {
+    const single = r.from === r.to;
+    const cls = [
+      'day',
+      single ? '' : 'range',
+      r.from <= today && today <= r.to ? 'today' : '',
+      r.to < today ? 'past' : '',
+      r.uid === mine ? 'mine' : '',
+    ].filter(Boolean).join(' ');
+    const label = single
+      ? [h('span', { class: 'day-name', text: dayName.format(parse(r.from)) }),
+        h('span', { class: 'day-date', text: dayDate.format(parse(r.from)) })]
+      : [h('span', {
+        class: 'day-range',
+        text: `${dayName.format(parse(r.from))} ${dayDate.format(parse(r.from))} – ${dayName.format(parse(r.to))} ${dayDate.format(parse(r.to))}`,
+      })];
+    return h('li', { class: cls },
+      ...label,
+      h('span', { class: `day-who ${r.entry ? '' : 'nobody'}` }, r.entry ? r.entry.username : 'nobody', r.entry && you(r.entry.user_id)),
+    );
+  });
   const thisWeek = isoDate(weekStart) === isoDate(mondayOf(new Date()));
   return [
     h('div', { class: 'page-head' },
@@ -121,12 +149,12 @@ function weekCard() {
         h('button', { class: 'btn btn-ghost btn-icon', type: 'button', 'aria-label': 'Previous week', onclick: () => shiftWeek(-1) },
           icon('chevronLeft')),
         h('button', {
-          class: 'btn btn-ghost label',
+          class: 'btn btn-ghost week-label',
           type: 'button',
           title: 'Back to this week',
           onclick: () => { weekStart = mondayOf(new Date()); refresh(); },
-          text: `Week ${isoWeek(weekStart)}`,
-        }),
+          text: `${dayDate.format(weekStart)} – ${dayDate.format(addDays(weekStart, 6))}`,
+        }, h('small', { text: ` Week ${isoWeek(weekStart)}` })),
         h('button', { class: 'btn btn-ghost btn-icon', type: 'button', 'aria-label': 'Next week', onclick: () => shiftWeek(1) },
           icon('chevronRight')),
       ),
@@ -135,7 +163,9 @@ function weekCard() {
   ];
 }
 
-// myShifts groups your upcoming dates into runs of consecutive days.
+// myShifts groups your upcoming dates into runs of consecutive days, then
+// splits off the one you're already in — listing it again under "Next
+// shifts" told people they hadn't started a shift they were already on.
 function myShifts() {
   const mine = data.upcoming.filter((e) => e.user_id === myID()).map((e) => e.date).sort();
   const runs = [];
@@ -145,16 +175,34 @@ function myShifts() {
     else runs.push({ from: date, to: date });
   }
   const fmt = (s) => `${dayName.format(parse(s))} ${dayDate.format(parse(s))}`;
-  return [
-    h('div', { class: 'page-head' }, h('h2', { text: 'Your next shifts' })),
+  const label = (r) => (r.from === r.to ? fmt(r.from) : `${fmt(r.from)} – ${fmt(r.to)}`);
+
+  const today = isoDate(new Date());
+  const current = runs[0] && runs[0].from <= today ? runs[0] : null;
+  const next = current ? runs.slice(1) : runs;
+
+  const currentCard = current ? [
+    h('div', { class: 'page-head' }, h('h2', { text: 'Current shift' })),
     h('div', { class: 'card' },
-      runs.length
-        ? h('ul', { class: 'shift-list' }, runs.slice(0, 8).map((r) =>
+      h('ul', { class: 'shift-list' }, h('li', {},
+        h('span', { text: label(current) }),
+        h('span', { class: 'muted', text: `ends in ${duration(addDays(parse(current.to), 1) - Date.now())}` })))),
+  ] : [];
+
+  return [
+    ...currentCard,
+    h('div', { class: 'page-head' }, h('h2', { text: current ? 'Next shifts' : 'Your next shifts' })),
+    h('div', { class: 'card' },
+      next.length
+        ? h('ul', { class: 'shift-list' }, next.slice(0, 8).map((r) =>
           h('li', {},
-            h('span', { text: r.from === r.to ? fmt(r.from) : `${fmt(r.from)} – ${fmt(r.to)}` }),
+            h('span', { text: label(r) }),
             h('span', { class: 'muted', text: days(r) })),
         ))
-        : h('div', { class: 'empty', text: 'Nothing scheduled in the next 60 days.' })),
+        : h('div', {
+          class: 'empty',
+          text: current ? 'Nothing else scheduled in the next 60 days.' : 'Nothing scheduled in the next 60 days.',
+        })),
   ];
 }
 

internal/web/static/js/queue.js

@@ -2,7 +2,7 @@
 
 import * as api from './api.js';
 import { h, clear, badge, emptyState, spinner } from './ui.js';
-import { age, until, isFuture, severityClass, labelSummary, teamColorClass } from './format.js';
+import { ago, until, isFuture, severityClass, labelSummary, teamColorClass } from './format.js';
 import { state, myID, setSelectedTeam, onTeamChange } from './state.js';
 import * as onboarding from './onboarding.js';
 import { navigate } from './app.js';
@@ -18,12 +18,12 @@ const FILTERS = [
 ];
 
 const EMPTY = {
-  open: ['All clear', 'Nothing open right now.'],
-  triggered: ['Nothing triggered', 'Every open incident has been acknowledged.'],
-  acknowledged: ['Nothing acknowledged', 'No one is working an incident right now.'],
-  snoozed: ['Nothing snoozed', 'Snoozed incidents show up here until the snooze runs out.'],
-  resolved: ['Nothing resolved', 'Resolved incidents are archived after a while.'],
-  archived: ['Nothing archived', ''],
+  open: ['All clear', 'Nothing open right now.', 'checkCircle'],
+  triggered: ['Nothing triggered', 'Every open incident has been acknowledged.', 'checkCircle'],
+  acknowledged: ['Nothing acknowledged', 'No one is working an incident right now.', 'checkCircle'],
+  snoozed: ['Nothing snoozed', 'Snoozed incidents show up here until the snooze runs out.', 'clock'],
+  resolved: ['Nothing resolved', 'Resolved incidents are archived after a while.', null],
+  archived: ['Nothing archived', 'Resolved incidents land here once archived.', 'archive'],
 };
 
 onboarding.onRerender(() => renderList());
@@ -87,6 +87,10 @@ export async function refresh({ fresh = false } = {}) {
     if (requested !== filter) return;
     error = err.message;
   }
+  // state.open (what the chip counts read) has just been refreshed too, by
+  // whichever caller updated it before calling here — app.js's poll, or the
+  // `cached` branch above.
+  renderChips();
   renderList();
 }
 
@@ -101,18 +105,32 @@ function setFilter(id) {
   refresh({ fresh: true });
 }
 
+// Counts for the three chips derivable from the open list already fetched
+// for the badges — Snoozed/Resolved/Archived would need a request of their
+// own, so those chips stay count-less for now.
+function chipCount(id) {
+  const open = state.selectedTeamID == null
+    ? state.open
+    : state.open.filter((i) => i.team_id === state.selectedTeamID);
+  if (id === 'open') return open.length;
+  if (id === 'triggered') return open.filter((i) => i.status === 'triggered').length;
+  if (id === 'acknowledged') return open.filter((i) => i.status === 'acknowledged').length;
+  return null;
+}
+
 function renderChips() {
   const el = document.getElementById('queue-filters');
-  const chips = FILTERS.map((f) =>
-    h('button', {
+  const chips = FILTERS.map((f) => {
+    const count = chipCount(f.id);
+    return h('button', {
       class: 'chip',
       type: 'button',
       role: 'tab',
       'aria-selected': String(f.id === filter),
       onclick: () => setFilter(f.id),
       text: f.label,
-    }),
-  );
+    }, count != null && h('span', { class: 'count', text: String(count) }));
+  });
 
   // Somebody in one team has nothing to choose between, so the row of team
   // chips appears only when there is more than one. The default is all of
@@ -140,6 +158,12 @@ function renderChips() {
     }
   }
 
+  // A scroll hint for the phone-width row, where the chips can run off the
+  // right edge with nothing to suggest there's more; the desktop sidebar
+  // wraps instead of scrolling (see .pane-list .chips), so this fades out
+  // there via CSS rather than being left out here.
+  chips.push(h('span', { class: 'chips-fade', 'aria-hidden': 'true' }));
+
   clear(el, chips);
 }
 
@@ -155,8 +179,8 @@ function renderList() {
     return;
   }
   if (!items.length) {
-    const [title, text] = EMPTY[filter];
-    clear(el, checklist, emptyState(title, text, filter === 'open' ? 'checkCircle' : null));
+    const [title, text, iconName] = EMPTY[filter];
+    clear(el, checklist, emptyState(title, text, iconName));
     return;
   }
   clear(el,
@@ -201,9 +225,12 @@ function row(inc, index) {
     dataset: { index: String(index) },
   },
     h('div', { class: 'row-title', text: inc.title }),
-    h('div', { class: 'row-age', title: inc.triggered_at, text: age(inc.triggered_at) }),
+    h('div', { class: 'row-age', title: inc.triggered_at, text: `Triggered ${ago(inc.triggered_at)}` }),
     h('div', { class: 'row-meta' },
       status,
+      // The left-border colour alone doesn't say what it means; spell it out
+      // too, same badge the incident detail page uses for severity.
+      inc.severity && badge(inc.severity, `plain ${severityClass(inc.severity)}`),
       assignee,
       team,
       labels && h('span', { class: 'labels', text: labels }),

internal/web/static/js/stats.js

@@ -80,7 +80,7 @@ function render() {
   if (error && !data) body = h('div', { class: 'load-error', text: error });
   else if (!data) body = spinner();
   else if (!data.incidents.total && !data.byHour.some((x) => x.count)) {
-    body = emptyState('No data in this range', '', 'chart');
+    body = emptyState('No data in this range', 'Nothing happened in this window.', 'chart');
   } else {
     body = h('div', { class: 'stats' },
       error && h('div', { class: 'load-error', text: `Showing older data: ${error}` }),

internal/web/static/js/team.js

@@ -182,7 +182,7 @@ function subnav() {
 // to the global team selector in the nav, which is what onTeamChange above
 // reacts to.
 function teamPicker() {
-  return h('div', { class: 'card' }, h('h2', { text: data.team.name }));
+  return h('h1', { class: 'detail-title', text: data.team.name });
 }
 
 // --- overview --------------------------------------------------------------
@@ -204,18 +204,18 @@ function overview() {
     menuCard('/team/rota', 'Rota', null,
       onToday ? `${onToday.username} is on call today.` : 'Nobody is on call today.'),
     menuCard('/team/members', 'Members', (data.members || []).length,
-      owners === 1 ? 'One owner.' : `${owners} owners.`),
+      owners === 1 ? '1 owner, full access.' : `${owners} owners, full access.`),
     menuCard('/team/escalation', 'Escalation', levels || null,
       levels
-        ? `${levels === 1 ? 'One level' : `${levels} levels`}${data.escalation.fallback_topic ? ', then a fallback topic.' : '.'}`
+        ? `${levels} escalation level${levels === 1 ? '' : 's'}${data.escalation.fallback_topic ? ', then a fallback topic.' : '.'}`
         : 'No ladder — nobody but the first person is woken.'),
     menuCard('/team/sources', 'Alert sources', keys || null,
       keys
-        ? (unused ? `${unused} of them never used.` : 'All in use.')
+        ? (unused ? `${unused} of ${keys} never used.` : `${keys} key${keys === 1 ? '' : 's'}, all used recently.`)
         : 'No key yet, so nothing can reach this team.'),
     menuCard('/team/deadman', 'Dead man’s switches', switches || null,
       switches
-        ? (dead ? `${dead} of them silent.` : 'All quiet, as they should be.')
+        ? (dead ? `${dead} of ${switches} silent.` : [icon('checkCircle', 'icon overview-note-icon'), ' Dead man’s switch: healthy.'])
         : 'Nothing watched.'),
   );
 }

internal/web/static/js/teamselector.js

@@ -5,7 +5,7 @@
 // between, the same rule every other team-aware control in this app follows;
 // see state.js's currentTeam() for why nobody with just one ever has to.
 
-import { h, openSheet, closeSheet } from './ui.js';
+import { h, icon, openSheet, closeSheet } from './ui.js';
 import { state, currentTeam, setSelectedTeam, onTeamChange } from './state.js';
 import { teamColorClass } from './format.js';
 
@@ -36,6 +36,9 @@ export function render() {
     btn.replaceChildren(
       h('span', { class: dotClass }),
       h('span', { class: 'team-selector-label', text: label }),
+      // Without this the pill reads as a tag rather than something you can
+      // open; the chevron is the only thing that says "dropdown" here.
+      icon('chevronDown', 'icon team-selector-chevron'),
     );
   }
 }

internal/web/static/js/ui.js

@@ -52,6 +52,7 @@ const ICONS = {
   trash: ['M4 7h16', 'M9 7V4h6v3', 'M6 7l1 13h10l1-13'],
   chevronLeft: ['M15 18l-6-6 6-6'],
   chevronRight: ['M9 6l6 6-6 6'],
+  chevronDown: ['M6 9l6 6 6-6'],
   external: ['M14 4h6v6', 'M20 4l-9 9', 'M18 14v6H4V6h6'],
   logout: ['M15 4h4v16h-4', 'M10 17l5-5-5-5', 'M15 12H4'],
 };
@@ -186,12 +187,16 @@ export function labelChip(k, v) {
 // One entry in a section's overview: a card that is a link, carrying the count
 // only that section can state. Both the Admin tab and the Team tab open on one
 // of these menus, and a menu item is a shape rather than a page's own idea.
+// note is usually just a string, but can be an array of children instead —
+// deadman's "healthy" note below pairs text with a status icon.
 export function menuCard(href, label, count, note) {
   return h('a', { class: 'card overview-item', href },
     h('div', { class: 'overview-head' },
       h('strong', { text: label }),
       count != null && h('span', { class: 'overview-count', text: String(count) })),
-    h('p', { class: 'muted small', text: note }),
+    typeof note === 'string'
+      ? h('p', { class: 'muted small', text: note })
+      : h('p', { class: 'muted small' }, note),
   );
 }