I've been struggling with a bizarre issue while using ArgoCD and Crossplane for a week now. Despite ArgoCD showing resources as "Healthy" and "Synced," Crossplane is failing to provision AWS resources, generating 400 errors. Meanwhile, the ArgoCD dashboard insists everything is fine, which is incredibly frustrating. I'm seeing problems with Lambda functions, stuck RDS instances, and IAM roles not being created, all while ArgoCD keeps up the facade. No matter how much I search online, I can't find anyone else reporting this, which makes me feel like I'm in isolation. I suspect the issue lies in how ArgoCD interprets the health check responses, since they prioritize `Ready: True` over `Synced: False`. I managed to fix it by changing the order of the checks, but I'm left wondering: is nobody else experiencing this? Are people not using health checks with Crossplane, relying on direct AWS monitoring instead? Or did I somehow misconfigure everything?
2 Answers
This seems to stem from a fundamental misunderstanding of GitOps. ArgoCD is functioning as it should; it aligns the cluster state with your repo. If Crossplane is failing afterward, that's a separate issue. You need monitoring systems like Datadog or Grafana to manage actual health checks, not just rely on ArgoCD's report.
Glad you found a workaround! But it's a bummer that you put it behind a Medium paywall. Maybe consider sharing it publicly elsewhere?
100% agree, Medium can be tricky with those member-only stories!
I avoid Medium too; if it's member only, I'm out.