Application Health Degradation Undetected by ArgoCD
warningArgoCD reports applications as 'Healthy' even when serving errors or experiencing high latency because built-in health checks only verify Kubernetes resource readiness, not actual application behavior.
Applications show Healthy status in ArgoCD but external monitoring (Prometheus, Datadog) shows elevated error rates or latency. Detect by comparing argocd_app_health_status (Healthy) with application-specific metrics showing degraded performance. Gap between ArgoCD health and actual service health indicates insufficient health checks.
Integrate ArgoCD with application performance monitoring: use Argo Rollouts with analysis templates that query Prometheus/Datadog metrics for error rates, latency percentiles, and business KPIs. Define custom health checks in argocd-cm for critical services that validate application-specific endpoints or dependencies. Configure notification triggers based on metric thresholds, not just sync status.