Gitlab pipelines failures

[fitillo]

We've been using fluxcd for a long time now. We've recently upgraded to v2.1.2 and also deployed several new apps. Installation was done as mentioned here https://fluxcd.io/flux/cmd/flux_bootstrap_gitlab/ and we follow a mono repo approach where we keep different environments in different folders.

When we do a deployment, we commit to flux repo to main branch and then check for commit status like this:

STATUS=$(curl -s --header "PRIVATE-TOKEN:${SOME_TOKEN}" "https://gitlab.com/api/v4/projects/fluxcd_repo/repository/commits/${HASH}" | jq -r .status)

The pipeline created for each commit always has the same amount of jobs with names like "kustomization/flux-system/some_id". Recently, the job for one particular ID is failing while all the other ones succeed. I've tried matching this id to kustomizations with no luck and also tried to search in flux pods logs, but couldn't find anything. I haven't found any clues in the docs either. Not sure if someone could help to narrow down the issue here.

[fitillo]

Ok, I found the issue myself. The ids relate to the UUID found in the provider configured by fluxcd. To narrow down the issue:

1. kubectl get providers -n flux-system
2. kubectl describe provider ${YOUR_PROVIDER_HERE} -n flux-system
3. Go and check logs for kustomize-controller for that particular cluster