fix(cilium): size the endpoint policy map and alert before it fills

Closes #451. Closes #454.

Both issues are the same fault: a pod's dedicated cilium_policy_v3_<id>
BPF map ran out of space, endpoint regeneration failed with "no space
left on device", the CNI ADD call never returned, and the pod sat in
sandbox creation. For paperless — replicas: 1 with strategy Recreate —
that was a 41-minute outage on 2026-07-30.

It is not over. Prometheus (metrics landed via #452) puts
cilium_bpf_map_pressure{map_name="cilium_policy_v3_*"} at 0.867 on all
three nodes right now, and max_over_time(...[30d]) at 1.03 — the map has
been driven past capacity. Loki has the overflow recurring as recently
as 2026-08-12 17:46Z, hitting authentik and esphome endpoints.

Three changes:

- bpf.policyMapMax 16384 (chart default) -> 65536, matching the cluster's
  existing bpf-lb-map-max. Takes steady-state pressure from ~87% to ~22%.
  The map is sized at creation, so this only takes effect as agents
  restart.
- The chart range ">=1.19.0" is what let Flux take 1.19.6 -> 1.20.0
  unattended 45 minutes before the first outage. Narrowed to
  ">=1.20.0 <1.21.0" so a CNI minor is a reviewed diff, not a surprise.
- A PrometheusRule on cilium_bpf_map_pressure: warning over 80% for 15m,
  critical at 100%. The metric was already scraped but nothing alerted on
  it, so the first signal was a downstream replica mismatch.

Not addressed: why the busiest endpoint needs ~14k entries. Only 178
CiliumIdentities exist and no CiliumNetworkPolicy declares more than 9
ports, so the entry count is not obviously explained by policy shape.
Establishing that needs `cilium-dbg bpf policy get` on an agent.

MERGING THIS ROLLS THE CNI. The DaemonSet uses
updateStrategy.rollingUpdate.maxUnavailable: 2, so two of three agents
restart at once.

Co-Authored-By: Claude Bot <claude-bot@renner.uno>

kubernetes/apps/monitoring-rules/kustomization.yaml

@@ -4,6 +4,7 @@ kind: Kustomization
 # come from the kube-prometheus-stack chart. Flux applies this Kustomization
 # only after `monitoring` is Ready (chart installed → CRDs present).
 resources:
+  - prometheusrule-cilium.yaml
   - prometheusrule-hetzner-backup.yaml
   - prometheusrule-loki.yaml
   - prometheusrule-netbird-availability.yaml

kubernetes/apps/monitoring-rules/prometheusrule-cilium.yaml

@@ -0,0 +1,37 @@
+apiVersion: monitoring.coreos.com/v1
+kind: PrometheusRule
+metadata:
+  name: cilium
+  namespace: monitoring
+  labels:
+    # kube-prometheus-stack's Prometheus resource selects on this label;
+    # without it the rule is created but never loaded into Prometheus.
+    release: kube-prometheus-stack
+spec:
+  groups:
+    - name: cilium.rules
+      rules:
+        - alert: CiliumBPFMapPressureHigh
+          # cilium_bpf_map_pressure is used/capacity for the fullest map of
+          # that name on the agent. A full map is not a slow path — it is a
+          # hard failure: the agent cannot program the datapath, so endpoint
+          # regeneration fails and every new pod on the node hangs in sandbox
+          # creation until something frees space.
+          expr: cilium_bpf_map_pressure > 0.8
+          for: 15m
+          labels:
+            severity: warning
+          annotations:
+            summary: "Cilium BPF map {{ $labels.map_name }} is {{ $value | humanizePercentage }} full on {{ $labels.node }}"
+            description: "The BPF map {{ $labels.map_name }} on node {{ $labels.node }} has been over 80% full for 15 minutes. At 100% the Cilium agent can no longer program the datapath and new pods on this node fail sandbox creation. Raise the corresponding size in kubernetes/infrastructure/cilium/helmrelease.yaml (values.bpf) — the map is sized at creation, so the change needs an agent restart."
+
+        - alert: CiliumBPFMapFull
+          # Pressure reaches (and can exceed) 1.0 when the agent wants more
+          # entries than the map holds. Pods are already failing to start.
+          expr: cilium_bpf_map_pressure >= 1
+          for: 0m
+          labels:
+            severity: critical
+          annotations:
+            summary: "Cilium BPF map {{ $labels.map_name }} is full on {{ $labels.node }}"
+            description: "The BPF map {{ $labels.map_name }} on node {{ $labels.node }} is at or over capacity. Endpoint regeneration is failing with 'no space left on device' and pods scheduled to this node will hang in ContainerCreating. Check `{namespace=\"kube-system\",pod=~\"cilium-.+\"} |= \"no space left on device\"` in Loki for the affected endpoints."

kubernetes/infrastructure/cilium/helmrelease.yaml

@@ -17,7 +17,11 @@ spec:
   chart:
     spec:
       chart: cilium
-      version: ">=1.19.0"
+      # Patch-only range: Flux reconciles every 30m and would otherwise take a
+      # CNI minor or major unattended. Widening this is a deliberate, reviewed
+      # act — a minor bump changes datapath behaviour and rolls 2 of 3 agents
+      # at once.
+      version: ">=1.20.0 <1.21.0"
       sourceRef:
         kind: HelmRepository
         name: cilium
@@ -46,6 +50,14 @@ spec:
       serviceMonitor:
         enabled: true
     bpf:
+      # Size of each endpoint's own cilium_policy_v3_<id> map, holding one
+      # entry per (identity, direction, protocol, port) the endpoint is
+      # allowed. When it fills, endpoint regeneration fails with "no space
+      # left on device", the CNI ADD call never returns, and the pod is stuck
+      # in sandbox creation — so this must stay comfortably above the busiest
+      # endpoint's entry count, tracked by cilium_bpf_map_pressure.
+      # Takes effect on agent restart; the map is sized at creation.
+      policyMapMax: 65536
       # Let pods reach the cluster's own LoadBalancer external IPs; Cilium
       # redirects the socket directly to the backend pod instead of sending
       # traffic out to the VIP and back (which fails with L2 announcements).