infra: export appset and mirror alert polish

StatefulSet/authentik-postgres has been eternally OutOfSync since ~Sprint 65
even though 'kubectl diff --server-side --field-manager=argocd-controller'
shows zero real change. The STS was created via ServerSideApply, so the live
object carries apiVersion/kind inside volumeClaimTemplates[]; git omitting
them makes ArgoCD's normalized diff disagree forever. Declare them in git.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

README.md

@@ -2,6 +2,22 @@
 
 Infrastructure manifests for ArgoCD. An `ApplicationSet` in `argocd` namespace watches the `apps/*` directories in this repo and creates one `Application` per subdir (prefixed `infra-<name>`).
 
+## Root GitOps ApplicationSet
+
+`argocd/applicationset-bluejay-infra.yaml` is the root of this GitOps tree, but
+it is **NOT self-managed** by ArgoCD. Apply it manually when the root generator
+or sync policy changes:
+
+```bash
+kubectl -n argocd apply -f argocd/applicationset-bluejay-infra.yaml
+```
+
+Keep the per-StatefulSet `ignoreDifferences` entries in that file synced with
+the live ApplicationSet. They intentionally cover `zabbix-postgres`,
+`guac-mysql`, `matrix-postgres`, and `authentik-postgres` so ArgoCD does not
+loop forever on server-side-apply `volumeClaimTemplates` status drift. Every new
+StatefulSet with `volumeClaimTemplates` needs its own entry appended.
+
 ## Adding a new service to the cluster
 
 Follow these steps in order. **Step 1 must run before step 3** — if you skip it, cert-manager HTTP-01 will silently fail for ~2h per cert (exponential backoff) until someone diagnoses the DNS.

apps/authentik/authentik.yaml

@@ -113,7 +113,12 @@ spec:
             - name: pgdata
               mountPath: /var/lib/postgresql/data
   volumeClaimTemplates:
-    - metadata:
+    # apiVersion/kind included deliberately: this STS was created via ArgoCD ServerSideApply,
+    # so the live object carries PVC TypeMeta inside volumeClaimTemplates; omitting it here
+    # leaves the app eternally OutOfSync even though kubectl SSA dry-run shows no change.
+    - apiVersion: v1
+      kind: PersistentVolumeClaim
+      metadata:
         name: pgdata
       spec:
         storageClassName: longhorn

apps/fc-messageboard/fc-messageboard.yaml

@@ -42,7 +42,7 @@ spec:
         app: messageboard-web
       annotations:
         fc.flowercore.io/healthz-anon: "true"
-        fc.flowercore.io/probe-path: "/healthz"
+        fc.flowercore.io/probe-path: "/health"
         prometheus.io/scrape: "true"
         prometheus.io/port: "8080"
         prometheus.io/path: "/metrics/prometheus"

apps/fc-ttsreader/fc-ttsreader.yaml

@@ -525,7 +525,7 @@ spec:
         app.kubernetes.io/part-of: flowercore
       annotations:
         fc.flowercore.io/healthz-anon: "true"
-        fc.flowercore.io/probe-path: "/healthz"
+        fc.flowercore.io/probe-path: "/health"
         prometheus.io/scrape: "true"
         prometheus.io/port: "5217"
         prometheus.io/path: "/metrics"

apps/fc-updater/fc-updater.yaml

@@ -54,7 +54,7 @@ spec:
     metadata:
       annotations:
         fc.flowercore.io/healthz-anon: "true"
-        fc.flowercore.io/probe-path: "/healthz"
+        fc.flowercore.io/probe-path: "/"
       labels:
         app: updatecenter-web
     spec:

apps/github-runner/README.md

@@ -24,6 +24,12 @@ original Longhorn ReadWriteOnce NuGet PVC. Every other repo-scoped runner uses
 two replicas with per-pod `emptyDir` caches. That is the safe backlog-drain
 strategy: no two pods share one RWO PVC.
 
+Ephemeral runner pods are expected to register, run one job, deregister, and
+exit so the Deployment starts a fresh pod for the next registration token. A
+small amount of exit-1/restart churn from token-expiry or no-work windows is
+accepted operational noise as long as jobs are not stuck queued and the
+repo-scoped runner-offline alerts stay quiet.
+
 Sprint 32 final long-tail wave adds 16 two-replica Deployments:
 `FlowerCore.Knowledge`, `FlowerCore.LlmBridge`, `FlowerCore.Media`,
 `FlowerCore.Presentations`, `FlowerCore.RemoteDesktop`, `FlowerCore.DNS`,

apps/monitoring/noc-monitoring.yaml

@@ -843,7 +843,9 @@ data:
         rules:
           - alert: PiManagerDown
             expr: up{job="pimanager-app"} == 0
-            for: 3m
+            # Sprint 67: delayed behind NodeDown's critical page so a powered-off
+            # Pi does not create the first duplicate page for the same host.
+            for: 8m
             labels:
               severity: warning
             annotations:
@@ -1242,6 +1244,58 @@ data:
               summary: "Marquee animation duration drifting > 10% on {{ $labels.renderer }} ({{ $labels.phase }})"
               description: "Median observed cycle duration deviates from target DurationMs by >10%. Could indicate browser tab throttling, GPU pressure, or phase-advancement bug."
 
+      # ============================================================
+      # Update Center public-edge probes
+      # Live-mirrored from FlowerCore.Notes/scripts/monitoring/alerts.yml.
+      # This K8s ConfigMap is the future migration target; live Prometheus
+      # still reads the canonical Notes file from noc1 Podman.
+      # ============================================================
+      - name: update_center
+        rules:
+          # Critical only when the edge is genuinely unreachable. A Cloudflare
+          # HTTP 429 means the prober hit a rate-limit, not that real clients
+          # are down, so the warning rule below owns that signal.
+          - alert: UpdateCenterPublicEdgeDown
+            expr: |
+              (probe_success{job="probe-update-center-public-edge"} == 0)
+              unless on(instance)
+              (probe_http_status_code{job="probe-update-center-public-edge"} == 429)
+            for: 10m
+            labels:
+              severity: critical
+              service: update-center
+              alert_channel: irc
+            annotations:
+              summary: "Update Center public edge probe failed for {{ $labels.instance }}"
+              description: >-
+                The external probe for {{ $labels.instance }} failed for 10 minutes with a
+                non-2xx status that is not a rate-limit. Public Update Center clients may be
+                unable to fetch manifest schema metadata through Cloudflare.
+              runbook: >-
+                1. curl -sk https://{{ $labels.instance }}/api/v1/manifests/_schema
+                2. Verify Cloudflare DNS record is proxied and targets the current public edge IP
+                3. kubectl -n fc-updater get ingressroute updatecenter-web-public secret cf-origin-flowercore-io
+                4. Check Traefik logs for Method() or TLS secret errors
+
+          - alert: UpdateCenterPublicEdgeRateLimited
+            expr: probe_http_status_code{job="probe-update-center-public-edge"} == 429
+            for: 15m
+            labels:
+              severity: warning
+              service: update-center
+              alert_channel: irc
+            annotations:
+              summary: "Cloudflare is rate-limiting (HTTP 429) the public-edge probe for {{ $labels.instance }}"
+              description: >-
+                The blackbox prober receives HTTP 429 from Cloudflare for {{ $labels.instance }}
+                while the origin is healthy. This is a Cloudflare rate-limit / WAF condition on
+                the public hostname, not an outage.
+              runbook: >-
+                1. curl -sk https://{{ $labels.instance }}/api/v1/manifests/_schema (expect 200 from a normal client)
+                2. Review Cloudflare rate-limit / WAF rules for the hostname; the 5m-cadence prober is tripping a 429
+                3. Add a Cloudflare rate-limit exception for the prober source IP or the /api/v1/manifests/_schema path
+                4. Confirm whether the singular host update.flowercore.io is still required, or only updates.flowercore.io
+
 # =============================================================================
 # ConfigMap: Blackbox Exporter Configuration
 # =============================================================================

argocd/applicationset-bluejay-infra.yaml

@@ -0,0 +1,74 @@
+apiVersion: argoproj.io/v1alpha1
+kind: ApplicationSet
+metadata:
+  annotations:
+    argocd.argoproj.io/refresh: "true"
+  name: bluejay-infra
+  namespace: argocd
+spec:
+  generators:
+  - git:
+      directories:
+      - path: apps/*
+      repoURL: http://gitea-clusterip.gitea.svc:3000/bluejay/bluejay-infra.git
+      revision: main
+      template:
+        metadata: {}
+        spec:
+          destination: {}
+          project: ""
+  goTemplate: true
+  goTemplateOptions:
+  - missingkey=error
+  template:
+    metadata:
+      name: infra-{{.path.basename}}
+    spec:
+      destination:
+        server: https://kubernetes.default.svc
+      ignoreDifferences:
+      - group: apps
+        jqPathExpressions:
+        - .spec.volumeClaimTemplates[]?.status
+        jsonPointers:
+        - /spec/volumeClaimTemplates
+        kind: StatefulSet
+        name: zabbix-postgres
+        namespace: zabbix
+      - group: apps
+        jqPathExpressions:
+        - .spec.volumeClaimTemplates[]?.status
+        jsonPointers:
+        - /spec/volumeClaimTemplates
+        kind: StatefulSet
+        name: guac-mysql
+        namespace: guacamole
+      - group: apps
+        jqPathExpressions:
+        - .spec.volumeClaimTemplates[]?.status
+        jsonPointers:
+        - /spec/volumeClaimTemplates
+        kind: StatefulSet
+        name: matrix-postgres
+        namespace: matrix
+      - group: apps
+        jqPathExpressions:
+        - .spec.volumeClaimTemplates[]?.status
+        jsonPointers:
+        - /spec/volumeClaimTemplates
+        kind: StatefulSet
+        name: authentik-postgres
+        namespace: authentik
+      project: default
+      source:
+        path: '{{.path.path}}'
+        repoURL: http://gitea-clusterip.gitea.svc:3000/bluejay/bluejay-infra.git
+        targetRevision: main
+      syncPolicy:
+        automated:
+          prune: true
+          selfHeal: true
+        syncOptions:
+        - CreateNamespace=true
+        - ServerSideApply=true
+        - RespectIgnoreDifferences=true

tests/bluejay-infra-lint/FleetManifestLintTests.cs

@@ -87,10 +87,10 @@ public sealed class FleetManifestLintTests
             ["fc-devicemgmt"] = ("fc-devicemgmt-web", "/healthz"),
             ["fc-library"] = ("library-web", "/health"),
             ["fc-llm-bridge"] = ("fc-llm-bridge", "/healthz"),
-            ["fc-messageboard"] = ("messageboard-web", "/healthz"),
+            ["fc-messageboard"] = ("messageboard-web", "/health"),
             ["fc-retail"] = ("retail-web", "/healthz"),
-            ["fc-ttsreader"] = ("ttsreader-web", "/healthz"),
+            ["fc-ttsreader"] = ("ttsreader-web", "/health"),
-            ["fc-updater"] = ("updatecenter-web", "/healthz"),
+            ["fc-updater"] = ("updatecenter-web", "/"),
             ["knowledge"] = ("knowledge-web", "/healthz"),
             ["telephony"] = ("telephony-web", "/health"),
             ["worldbuilder"] = ("worldbuilder-web", "/healthz"),
@@ -468,6 +468,99 @@ public sealed class FleetManifestLintTests
         monitoring.Should().Contain("dedicated LinuxRunnerOffline/MacMiniRunnerOffline alerts");
     }
 
+    [Fact]
+    public void GithubRunnerReadme_DocumentsAcceptedEphemeralExitChurn()
+    {
+        var readme = File.ReadAllText(Path.Combine(Inventory.BluejayRoot, "apps", "github-runner", "README.md"));
+
+        readme.Should().Contain("Ephemeral runner pods");
+        readme.Should().Contain("exit-1/restart churn");
+        readme.Should().Contain("accepted operational noise");
+        readme.Should().Contain("repo-scoped runner-offline alerts stay quiet");
+    }
+
+    [Fact]
+    public void Monitoring_PiManagerDownDelayAndUpdateCenterRateLimit_MatchCanonicalAlerts()
+    {
+        var notesAlerts = File.ReadAllText(Path.Combine(
+            Inventory.WorkspaceRoot,
+            "FlowerCore.Notes",
+            "scripts",
+            "monitoring",
+            "alerts.yml"));
+        var monitoring = File.ReadAllText(Path.Combine(Inventory.BluejayRoot, "apps", "monitoring", "noc-monitoring.yaml"));
+
+        notesAlerts.Should().Contain("# Sprint 67: keep this warning behind NodeDown's 5m critical page");
+        notesAlerts.Should().Contain("- alert: PiManagerDown");
+        notesAlerts.Should().Contain("for: 8m");
+        monitoring.Should().Contain("# Sprint 67: delayed behind NodeDown's critical page");
+        monitoring.Should().Contain("- alert: PiManagerDown");
+        monitoring.Should().Contain("for: 8m");
+
+        notesAlerts.Should().Contain("- alert: UpdateCenterPublicEdgeRateLimited");
+        notesAlerts.Should().Contain("expr: probe_http_status_code{job=\"probe-update-center-public-edge\"} == 429");
+        notesAlerts.Should().Contain("for: 15m");
+        monitoring.Should().Contain("- alert: UpdateCenterPublicEdgeRateLimited");
+        monitoring.Should().Contain("expr: probe_http_status_code{job=\"probe-update-center-public-edge\"} == 429");
+        monitoring.Should().Contain("for: 15m");
+        monitoring.Should().Contain("severity: warning");
+    }
+
+    [Fact]
+    public void ApplicationSetExport_MustRemainManualRootOfGitOpsTree()
+    {
+        var readme = File.ReadAllText(Path.Combine(Inventory.BluejayRoot, "README.md"));
+        var appsetPath = Path.Combine(Inventory.BluejayRoot, "argocd", "applicationset-bluejay-infra.yaml");
+
+        File.Exists(appsetPath).Should().BeTrue();
+        var appset = File.ReadAllText(appsetPath);
+
+        appset.Should().Contain("kind: ApplicationSet");
+        appset.Should().Contain("name: bluejay-infra");
+        appset.Should().NotContain("\nstatus:");
+        appset.Should().NotContain("managedFields:");
+        readme.Should().Contain("root of this GitOps tree");
+        readme.Should().Contain("NOT self-managed");
+        readme.Should().Contain("kubectl -n argocd apply -f argocd/applicationset-bluejay-infra.yaml");
+    }
+
+    [Fact]
+    public void ApplicationSetExport_MustDiscoverAppsDirectoryOnMain()
+    {
+        var appset = File.ReadAllText(Path.Combine(Inventory.BluejayRoot, "argocd", "applicationset-bluejay-infra.yaml"));
+
+        appset.Should().Contain("path: apps/*");
+        appset.Should().Contain("revision: main");
+        appset.Should().Contain("repoURL: http://gitea-clusterip.gitea.svc:3000/bluejay/bluejay-infra.git");
+        appset.Should().Contain("path: '{{.path.path}}'");
+        appset.Should().Contain("targetRevision: main");
+        appset.Should().Contain("ServerSideApply=true");
+        appset.Should().Contain("RespectIgnoreDifferences=true");
+    }
+
+    [Fact]
+    public void ApplicationSetExport_MustPreserveStatefulSetIgnoreDifferences()
+    {
+        var appset = File.ReadAllText(Path.Combine(Inventory.BluejayRoot, "argocd", "applicationset-bluejay-infra.yaml"));
+
+        appset.Should().Contain("jsonPointers:");
+        appset.Should().Contain("- /spec/volumeClaimTemplates");
+        appset.Should().Contain(".spec.volumeClaimTemplates[]?.status");
+        Regex.Matches(appset, "kind: StatefulSet").Should().HaveCount(4);
+
+        foreach (var (name, ns) in new[]
+        {
+            ("zabbix-postgres", "zabbix"),
+            ("guac-mysql", "guacamole"),
+            ("matrix-postgres", "matrix"),
+            ("authentik-postgres", "authentik"),
+        })
+        {
+            appset.Should().Contain($"name: {name}");
+            appset.Should().Contain($"namespace: {ns}");
+        }
+    }
+
     [Fact]
     public void Monitoring_BlackboxTargetsForOidcSensitiveServices_MustUseAnonymousHealthRoutesWhenAvailable()
     {