Observability and Troubleshooting Kubernetes Workloads in Production

<p>Learn to see inside a Kubernetes-hosted application instead of guessing what's wrong. You'll instrument a real three-microservice Flask/PostgreSQL app with Prometheus metrics and OpenTelemetry traces, build a Grafana dashboard that actually answers "is this getting worse," and design Alertmanager routing and alert thresholds that page the right person only when it matters — backed by centralized logs in Loki. Then you'll apply a systematic observe-describe-isolate-fix-verify workflow to diagnose and fix three deliberately broken deployments: a bad image tag, a misconfigured readiness probe, and a NetworkPolicy silently blocking traffic — using kubectl describe, logs --previous, get events, and ephemeral debug containers.</p>

Tools: kubernetes, prometheus, grafana, loki, opentelemetry