[ECK] Clarification on APM Server resource management - What's handled by the operator vs manual configuration?

Kibana version: 8.14.3
Elasticsearch version: 8.14.3
APM Server version: 8.14.3
Original install method: Migrating from Helm chart to ECK operator (eck-stack helm chart version 0.12.1)
Fresh install or upgraded: Migration from classic helm deployment to ECK

Is there anything special in your setup?
Currently using classic helm deployment with following resources:

  • clusterrole/clusterrolebinding
  • configmap
  • deployment
  • service
  • serviceaccount
  • ingress
  • hpa

Planning to migrate to ECK-managed APM Server using eck-apm-server CRD.

Description of the problem including expected versus actual behavior:
Need clarification on which Kubernetes resources are automatically managed by the ECK operator versus which need manual management.

Current understanding is that ECK manages core resources through the CRD spec:

eck-apm-server:
  spec:
    config: {...}     # Replaces configmap
    podTemplate: {...} # Pod/deployment specs
    http: {...}       # Service configuration

Specific questions:

  1. Are RBAC resources (clusterrole/binding) automatically created and managed by the operator?
  2. Is HPA configuration supported through the CRD spec or needs separate resource?
  3. What resources still require manual creation/management?

Steps to reproduce:

  1. Have classic APM Server helm deployment with all resources listed above
  2. Migrating to ECK operator using eck-apm-server CRD
  3. Need to understand which resources to keep/create manually vs let operator manage

Would appreciate clarification on the operator's scope of resource management to ensure proper migration. Thank you!

From APM to Elastic Cloud on Kubernetes (ECK)

Removed agent-explorer, docker, elastic-stack-alerting, elastic-stack-monitoring, elastic-stack-security

Added docker, elastic-stack-alerting, elastic-stack-monitoring