Deploying this plugin leads to cluster status log spam

[baurmatt]

Reproduction commands:

git clone https://github.com/cloudnative-pg/cnpg-i-hello-world
cd cnpg-i-hello-world
kind create cluster --name cnpg-i-hello-world
kubectl apply --server-side -f https://github.com/cloudnative-pg/cloudnative-pg/releases/download/v1.25.0/cnpg-1.25.0.yaml
kubectl apply -f https://github.com/cert-manager/cert-manager/releases/download/v1.17.0/cert-manager.yaml
task local-kind-deploy
kubectl apply -f doc/examples/cluster-example.yaml

After the cluster is fully deployed, the following log output is written every second:

Operator

{
  "level": "info",
  "ts": "2025-02-06T13:02:49.941804502Z",
  "logger": "set_status_plugin_hook",
  "msg": "patching cluster status with the updated plugin statuses",
  "controller": "cluster",
  "controllerGroup": "postgresql.cnpg.io",
  "controllerKind": "Cluster",
  "Cluster": {
    "name": "cluster-example",
    "namespace": "default"
  },
  "namespace": "default",
  "name": "cluster-example",
  "reconcileID": "3932f6e6-f86c-4ef3-8443-dbc67e1a8cb3"
}

Plugin

{
  "level": "info",
  "ts": 1738846970.093195,
  "logger": "cnpg_i_example_lifecyle",
  "caller": "operator/status.go:57",
  "msg": "plugin is enabled, no action taken"
}

While I could imagine that the plugin output is there for showcasing, I'm quite sure that the operator output indicates a bug.

[ani1609]

@mnencia hey i would like to work on this

[mnencia]

@ani1609, welcome and thanks for offering your help! We really appreciate your willingness to contribute. If you’d like, we encourage you to share your analysis of the issue and feel free to open a pull request with any suggestions or fixes you might have.

[DevAnuragT]

Hey @mnencia,

I investigated the cause and wanted to share my findings before opening a PR.

Problem

The SetStatusInCluster method always returns a status update even if nothing has changed, which leads the operator to patch the status on every reconcile cycle. This causes the log spam described.

Code Location

The issue is in internal/operator/status.go:

return clusterstatus.NewSetStatusInClusterResponseBuilder().JSONStatusResponse(Status{Enabled: true})

This always returns a new status update. So the operator believes the plugin status has changed, even when it's not.

Solutions

Option 1 (Recommended for Hello World plugin):

Remove the SET_STATUS_IN_CLUSTER capability from internal/operator/impl.go. This simplifies the example and removes unnecessary logic for new developers.

Why this is recommended:

• The Hello World plugin is meant to be minimal and educational, so it’s better to keep the operator logic lean and clear.

• The plugin does not maintain or track any meaningful runtime state, so patching the status adds unnecessary complexity.

• Reduces confusion for new contributors exploring the plugin structure.

Option 2:

Fix the logic in SetStatusInCluster to compare the old and new plugin status and only respond with a patch if the status truly changes.

Additional Suggestions:

• Fix potential null pointer dereference in lifecycle.go

• Add error handling for ignored return values

• Properly initialize maps to avoid panics

I’m happy to submit a PR to address this. Let me know if you'd prefer the minimal change (Option 2), or a cleaner approach for the example (Option 1).