fix(observability): 🔇 silence managed-K8s false alerts + bump backup deadline to 4h

- Disable kubernetesSystemControllerManager, kubeScheduler, kubernetesSystemScheduler
  alert rules in template (unreachable on managed K8s)
- Bump forgejo s3 backup activeDeadlineSeconds 7200→14400 (2h→4h) in template;
  deadline hit Jun 20-21 on heavy sync

template/stacks/forgejo/forgejo-server/manifests/forgejo-s3-backup-cronjob.yaml

@@ -11,8 +11,8 @@ spec:
   startingDeadlineSeconds: 600 # 10 minutes
   jobTemplate:
     spec:
-      # 2h window: handles large incremental syncs after repo growth or OBS slowness; BackupJobTooSlow alert fires at 5m
-      activeDeadlineSeconds: 7200
+      # 4h window: bumped from 2h after Jun 20-21 deadline hit on heavy sync; BackupJobTooSlow alert fires at 5m
+      activeDeadlineSeconds: 14400
       backoffLimit: 2
       ttlSecondsAfterFinished: 259200 #
       template:

template/stacks/observability-client/vm-client-stack/values.yaml

@@ -201,13 +201,13 @@ defaultRules:
       create: true
       rules: {}
     kubernetesSystemControllerManager:
-      create: true
+      create: false
       rules: {}
     kubeScheduler:
-      create: true
+      create: false
       rules: {}
     kubernetesSystemScheduler:
-      create: true
+      create: false
       rules: {}
     kubeStateMetrics:
       create: true