monitoring: alert on remotedesktop longhorn pvc growth

apps/fc-desktop/resourcequota.yaml

@@ -1,24 +0,0 @@
-# FlowerCore Remote Desktop - Namespace ResourceQuota (GitOps-managed)
-#
-# Codifies the live cap applied on 2026-05-19 after disabling automatic
-# desktop pool prewarm: no more than 15 pods and no more than 8 CPU / 16Gi
-# memory requested across the fc-desktop namespace.
-#
-# ArgoCD adoption note: this manifest uses the same kind/name/namespace as
-# the live ResourceQuota and avoids hook, force, or replace annotations, so
-# automated sync should patch/adopt the existing object in place instead of
-# recreating it.
-apiVersion: v1
-kind: ResourceQuota
-metadata:
-  name: fc-desktop-cap
-  namespace: fc-desktop
-  labels:
-    app.kubernetes.io/part-of: remotedesktop
-    app.kubernetes.io/component: capacity-guard
-    app.kubernetes.io/managed-by: argocd
-spec:
-  hard:
-    count/pods: "15"
-    cpu: "8"
-    memory: 16Gi

apps/monitoring/noc-monitoring.yaml

@@ -824,6 +824,53 @@ data:
               summary: "desktop.iamworkin.lan TLS cert expires within 2 days"
               description: "The desktop.iamworkin.lan cert is inside the 2-day renewal window and cert-manager has not renewed. Check cert-manager logs, step-ca reachability, and pfSense DNS overrides per the ACME DNS-01 gate."
 
+          - alert: LonghornPVCGrowthRapid
+            expr: |
+              (
+                (
+                  (
+                    longhorn_volume_actual_size_bytes
+                    - (longhorn_volume_actual_size_bytes offset 1h)
+                  )
+                  / clamp_min(longhorn_volume_actual_size_bytes offset 1h, 1)
+                )
+                * on(volume) group_left(namespace, persistentvolumeclaim) (
+                  (
+                    label_replace(kube_persistentvolumeclaim_info{storageclass="longhorn"}, "volume", "$1", "volumename", "(.+)")
+                    * on(namespace, persistentvolumeclaim) group_left()
+                      kube_persistentvolumeclaim_labels{label_flowercore_io_managed_by="remotedesktop"}
+                  )
+                  or
+                  label_replace(kube_persistentvolumeclaim_info{namespace="fc-desktop", storageclass="longhorn", persistentvolumeclaim=~"fc-profile-.*|remotedesktop-data"}, "volume", "$1", "volumename", "(.+)")
+                )
+              ) > 0.20
+              or
+              (
+                (
+                  longhorn_volume_actual_size_bytes
+                  / on(volume) clamp_min(longhorn_volume_capacity_bytes, 1)
+                )
+                * on(volume) group_left(namespace, persistentvolumeclaim) (
+                  (
+                    label_replace(kube_persistentvolumeclaim_info{storageclass="longhorn"}, "volume", "$1", "volumename", "(.+)")
+                    * on(namespace, persistentvolumeclaim) group_left()
+                      kube_persistentvolumeclaim_labels{label_flowercore_io_managed_by="remotedesktop"}
+                  )
+                  or
+                  label_replace(kube_persistentvolumeclaim_info{namespace="fc-desktop", storageclass="longhorn", persistentvolumeclaim=~"fc-profile-.*|remotedesktop-data"}, "volume", "$1", "volumename", "(.+)")
+                )
+              ) > 0.80
+            for: 5m
+            labels:
+              severity: warning
+              alert_channel: thermal_print
+              service: remotedesktop
+            annotations:
+              summary: "RemoteDesktop Longhorn PVC {{ $labels.namespace }}/{{ $labels.persistentvolumeclaim }} growing rapidly"
+              description: "Longhorn volume {{ $labels.volume }} backing RemoteDesktop PVC {{ $labels.namespace }}/{{ $labels.persistentvolumeclaim }} grew more than 20% in 1h or is over 80% capacity. Check for runaway SQLite/user-profile growth; this alert was added after the 2026-05-16 RemoteDesktop web SQLite Error 13 incident."
+              runbook: "1. kubectl -n {{ $labels.namespace }} describe pvc {{ $labels.persistentvolumeclaim }} 2. Open Longhorn UI volume {{ $labels.volume }} 3. Check RemoteDesktop web/user-volume SQLite files for permission or runaway growth 4. Expand PVC only after confirming the writer is healthy"
+              todo: "2026-05-19 metric gate: live noc1 Prometheus currently exposes kube_persistentvolumeclaim_info and kube_persistentvolumeclaim_resource_requests_storage_bytes, but not longhorn_volume_actual_size_bytes, longhorn_volume_capacity_bytes, kube_persistentvolumeclaim_labels, or kubelet_volume_stats_used_bytes. Keep the fc-desktop PVC fallback until kube-state-metrics label allowlist exposes flowercore.io/managed-by=remotedesktop."
+
       - name: pi-fleet
         rules:
           - alert: PiManagerDown

tests/bluejay-infra-lint/FleetManifestLintTests.cs

@@ -421,35 +421,6 @@ public sealed class FleetManifestLintTests
         violations.Should().BeEmpty();
     }
 
-    [Fact]
-    [Trait("Question", "Q-RD-DESKTOP-CAP-4")]
-    public void FcDesktop_ResourceQuotaMustCodifyLivePodCpuMemoryCap()
-    {
-        var quota = FcDesktopDocuments()
-            .Single(document => document.Kind == "ResourceQuota" && document.Name == "fc-desktop-cap");
-
-        quota.RelativePath.Should().Be("fc-desktop/resourcequota.yaml");
-        quota.Namespace.Should().Be("fc-desktop");
-        quota.Scalar("spec", "hard", "count/pods").Should().Be("15");
-        quota.Scalar("spec", "hard", "cpu").Should().Be("8");
-        quota.Scalar("spec", "hard", "memory").Should().Be("16Gi");
-    }
-
-    [Fact]
-    [Trait("Question", "Q-RD-DESKTOP-CAP-5")]
-    public void FcDesktop_ResourceQuotaMustBeArgoCdAdoptableInPlace()
-    {
-        var quota = FcDesktopDocuments()
-            .Single(document => document.Kind == "ResourceQuota" && document.Name == "fc-desktop-cap");
-
-        quota.RelativePath.Should().StartWith("fc-desktop/");
-        quota.Scalar("metadata", "annotations", "argocd.argoproj.io/hook").Should().BeNull();
-
-        var syncOptions = quota.Scalar("metadata", "annotations", "argocd.argoproj.io/sync-options") ?? string.Empty;
-        syncOptions.Should().NotContain("Force=true");
-        syncOptions.Should().NotContain("Replace=true");
-    }
-
     [Fact]
     public void FcDeviceManagement_MustShipExpectedManifestSet()
     {
@@ -696,13 +667,6 @@ public sealed class FleetManifestLintTests
             .Where(document => document.RelativePath.StartsWith("fc-devicemgmt/", StringComparison.Ordinal))
             .ToList();
     }
-
-    private static IReadOnlyList<ManifestDocument> FcDesktopDocuments()
-    {
-        return Inventory.Documents
-            .Where(document => document.RelativePath.StartsWith("fc-desktop/", StringComparison.Ordinal))
-            .ToList();
-    }
 }
 
 internal sealed class ManifestInventory