Day-two problems on a running Argo CD: applications that stay OutOfSync or Progressing, deletions that hang, timeouts in manifest generation and lost admin access.
Diagnosing permanent drift: fields Kubernetes drops or normalizes, mutating webhooks and controllers, random values in Helm templates, HPA reordering, and instance label clashes with other tools.
Configuring ignoreDifferences per application or system-wide: JSON pointers, jq path expressions, managed fields managers, status field comparison and aggregated roles, without hiding real drift.
Finding why health never turns Healthy: Ingress without a load balancer address, StatefulSet rollouts, older SealedSecret CRDs, custom resources without health checks, and Lua checks that flap.
Handling deletions that hang: missing or corrupted source repositories, the resources finalizer, non-cascading deletes that orphan resources, and cleaning up safely.
Resolving reconciliation failures caused by slow manifest generation: repo-server timeouts, exec timeouts for Helm and Kustomize, parallelism limits, Git retries and adding repo-server capacity.
Getting back in after a lost admin password or a broken SSO configuration: resetting the bcrypt hash in argocd-secret, regenerating the initial password, token revocation, and fixing OIDC settings from the cluster.
Turning Argo CD metrics into useful alerts: OutOfSync and Degraded application counts from argocd_app_info, failed syncs, slow reconciliation, cluster connection loss, and routing alerts to teams by project.
Tell us what you're working on, whether it's a deployment, an audit, a security test or a cyber range. You'll speak with an engineer who can help you scope it.