Stop finding out about pod failures from your users.
This flow watches your Kubernetes pods in real-time, detects problems — CrashLoopBackOff, OOMKilled, failed scheduling, you name it — pulls the last 30 lines of logs from the
failing container, and drops a rich Slack message into your channel before you've even switched terminal tabs.
What's inside:
- Realtime pod watcher with optional label selector
- Automatic container log fetching from problematic pods
- Formatted Slack alerts with pod name, namespace, node, failure reason, and a log snippet
- 5-minute per-pod cooldown — no alert storms, no channel spam
- Graceful fallback: if log fetching fails (pod already evicted?), you still get the alert
Configure once: label selector, Slack channel ID, bot token. No YAML templating, no Helm charts, no AlertManager rule files.