✅ Q1 — ArgoCD shows “OutOfSync” but app is working fine. What do you check?

I compare desired vs live manifests in ArgoCD diff view. Often drift comes from fields modified by controllers (like replicas from HPA or injected sidecars). I check if those fields should be ignored. Then I configure ignoreDifferences in ArgoCD app spec. Never blindly force sync without reviewing diff.

✅ Q2 — ArgoCD sync fails but kubectl apply works manually — why?

Usually RBAC or service account permission issue for ArgoCD controller. ArgoCD applies using its own service account, not my kubectl identity. I check ArgoCD repo-server and application-controller logs. Then verify cluster role bindings.

✅ Q3 — Deployment keeps getting reverted after manual kubectl edit. Why?

Because ArgoCD is GitOps controller — Git is source of truth. Any manual cluster change is overwritten on next sync. Correct fix is update Git manifests and commit. Manual edits are anti-pattern in GitOps.

✅ Q4 — How do you safely deploy to production using ArgoCD?

Use manual sync instead of auto-sync for prod. Enable sync waves and hooks for ordering. Use health checks and sync windows. Combine with PR approval workflow in Git before merge.

✅ Q5 — ArgoCD shows app stuck in “Progressing” state — debug steps?

Check resource health details in UI. Then check Kubernetes events and pod status. ArgoCD health depends on readiness probes and rollout status. Often root cause is failing readiness or crashloop.

✅ Q6 — How do you manage secrets with ArgoCD?

Never store plain secrets in Git. Use Sealed Secrets, External Secrets Operator, or SOPS encryption. ArgoCD deploys encrypted manifests which are decrypted in cluster. Keeps Git safe.

✅ Q7 — Multi-cluster deployment using ArgoCD — how implemented?

Register multiple clusters in ArgoCD using cluster credentials. Create applications targeting different clusters. Use app-of-apps or ApplicationSet pattern for scale. Keeps single control plane for many clusters.

✅ Q8 — What is App of Apps pattern?

One parent ArgoCD application manages multiple child applications. Parent points to folder of app manifests. Useful for platform bootstrap and environment stacks. Makes cluster setup reproducible.

✅ Q9 — ArgoCD auto-sync caused production outage — how prevent?

Enable sync windows and manual approval for critical apps. Use progressive delivery (Argo Rollouts) instead of instant replace. Avoid auto-prune in sensitive workloads. Git merge controls are part of safety.

✅ Q10 — How do you handle config differences per environment in ArgoCD?

Use Kustomize overlays or Helm values per env. Same base manifests with env-specific patches. Separate folders or branches per environment. Avoid copy-paste manifests.

✅ Q11 — ArgoCD cannot pull from Git repo — checks?

Check repo credentials and token validity. Verify network access from repo-server pod. Check SSH key or PAT configuration. Look at repo-server logs for auth errors.

✅ Q12 — Difference between ArgoCD and Jenkins for deployment?

Jenkins pushes deployments via pipeline steps. ArgoCD pulls desired state from Git continuously. Jenkins = CI/push model. ArgoCD = GitOps/pull reconciliation model. Modern pattern is Jenkins builds, ArgoCD deploys.

✅ Q13 — How do you deploy Helm charts with ArgoCD?

ArgoCD has native Helm support. You point app to Helm chart repo and values file. ArgoCD renders templates and applies manifests. No Helm release state stored — ArgoCD manages resources.

✅ Q14 — ArgoCD pruning deleted a resource unexpectedly — why?

Prune removes resources not present in Git anymore. If manifest removed or renamed, ArgoCD deletes it. That’s expected GitOps behavior. Use prune carefully and review diffs.

✅ Q15 — How do you implement progressive delivery with ArgoCD?

Use Argo Rollouts instead of standard Deployment. Supports canary and blue-green strategies. ArgoCD deploys Rollout CRD. Traffic shifting controlled gradually with metrics checks.