Retry policy resolution when no tags found in database

[stefanprodan]

Changes:

• Add retry with backoff for listing tags to cope with race conditions at startup
• Add standard controller flag --requeue-dependency (default 30s)
• Requeue policies according to the flag when no tags are found in the database

Fix: #844

Release candidate

Images containing this fix:

• ghcr.io/fluxcd/image-reflector-controller:rc-0c2d58b3
• docker.io/fluxcd/image-reflector-controller:rc-0c2d58b3

Bootstrap patch:

apiVersion: kustomize.config.k8s.io/v1beta1
kind: Kustomization
resources:
  - gotk-components.yaml
  - gotk-sync.yaml
patches:
  - target:
      kind: Deployment
      name: image-reflector-controller
    patch: |
      - op: replace
        path: /spec/template/spec/containers/0/image
        value: ghcr.io/fluxcd/image-reflector-controller:rc-0c2d58b3

FluxInstance patch:

apiVersion: fluxcd.controlplane.io/v1
kind: FluxInstance
metadata:
  name: flux
  namespace: flux-system
spec:
  distribution:
    version: "2.7.x"
    registry: "ghcr.io/fluxcd"
  kustomize:
    patches:
      - target:
          kind: Deployment
          name: image-reflector-controller
        patch: |
          - op: replace
            path: /spec/template/spec/containers/0/image
            value: ghcr.io/fluxcd/image-reflector-controller:rc-0c2d58b3

[gustavobini]

I tested it out.

I deleted the image-reflector-controller pod multiple times. The repositories report "no tags in database" but immediately recover. The image policies do not even report the "no tags in database" error as before.

When deleting an ImageRepository, the ImagePolicy reports "no tags in database" as expected, and also immediately recovers when the ImageRepository is created.

It works very well, thank you!

[matheuscscp]

LGTM

[fluxcdbot]

Successfully created backport PR for release/v1.0.x:

• [release/v1.0.x] Retry policy resolution when no tags found in database #846