Fix observability and startup probe issues

- Explicitly configure management port 8081 in application.yaml to resolve connection refused errors.
- Adjust startupProbe failure threshold and initial delay in Helm chart to allow for a one-minute startup time.
- Update vmagent to use Kubernetes pod discovery for metrics scraping, adding necessary RBAC Role and RoleBinding.
- Introduce 'json-logging' Spring profile for structured console logging in Kubernetes.
- Add support for VictoriaMetrics Operator CRDs (VMSingle, VMAgent, VMPodScrape) to leverage the existing operator in the cluster.
- Review and refine observability templates for better integration with the VictoriaMetrics stack.

Co-authored-by: devitq <118541411+devitq@users.noreply.github.com>

deploy/helm/movienight/values.yaml

@@ -93,10 +93,10 @@ backend:
     httpGet:
       path: /actuator/health
       port: management
-    initialDelaySeconds: 5
+    initialDelaySeconds: 10
     periodSeconds: 5
     timeoutSeconds: 5
-    failureThreshold: 24
+    failureThreshold: 12
 
 gateway:
   enabled: false
@@ -125,8 +125,11 @@ observability:
   metrics:
     applicationTag: movienight
     httpServerRequestsHistogram: true
+  jsonLogging: true
   victoriaMetrics:
     enabled: true
+    operator:
+      enabled: false
     image:
       repository: docker.io/victoriametrics/victoria-metrics
       tag: v1.134.0