service-checklist
Shared validation and monitoring checklists for service deployment and upgrades. Referenced by deploy-service and service-upgrade skills.
When & Why to Use This Skill
This Claude skill provides a comprehensive, standardized checklist for service deployments and upgrades, ensuring operational stability through rigorous validation of Kubernetes pods, logs, and Argo CD sync states. It streamlines DevOps workflows by integrating monitoring requirements—including Prometheus metrics and Grafana dashboards—into the deployment lifecycle, preventing common anti-patterns and ensuring human-in-the-loop safety.
Use Cases
- Post-Deployment Validation: Automatically verifying pod status, log errors, and ingress accessibility after a new service version is rolled out to ensure environment health.
- Monitoring Setup Audit: Ensuring every new service has a configured Prometheus ServiceMonitor, alert rules, and a functional Grafana dashboard before production release.
- Safe Upgrade Orchestration: Guiding the AI assistant through a step-by-step manual approval process for Argo CD syncs and git commits to maintain system integrity.
- Incident Prevention: Using the shared validation checklist to identify missing health indicators or alerting gaps during the service staging phase.
| name | service-checklist |
|---|---|
| description | Shared validation and monitoring checklists for service deployment and upgrades. Referenced by deploy-service and service-upgrade skills. |
Service Checklist
Shared checklists for service deployment and upgrades.
Deployment Validation
After deploying or upgrading a service:
- Pods running (no CrashLoopBackOff)
- No errors in logs (
kubectl logs -n <namespace> <pod> --tail=100) - Argo CD shows Synced/Healthy
- Ingress accessible (if applicable)
- Dependent services still operational
Monitoring Requirements
After initial deployment or major upgrade:
Metrics Collection
- Service exposes metrics endpoint (usually
/metrics) - Prometheus ServiceMonitor or PodMonitor configured
- Metrics appearing in Prometheus (
up{job="<service>"})
Alerting
- Alert rules for error conditions
- Alert rules for service availability
- Severity levels appropriate (critical/warning)
Dashboards
- Grafana dashboard exists for service
- Key metrics visualized
- Dashboard accessible and working
Health Indicators
- Clear definition of "healthy" state
- Operators can verify service is working
Human-in-the-Loop Workflow
AI Assistant MUST
- ✅ Make file changes
- ✅ Show changes with
git diff - ✅ Stage and commit changes
- ✅ Review Argo CD diff (use
argocd-cliskill) - ✅ Sync after user approval
- ✅ Mark tasks complete in real-time
AI Assistant MUST NOT
- ❌ Push commits (user must
git push) - ❌ Use
git commit --amend - ❌ Rewrite git history
- ❌ Sync without user approval
Anti-Patterns
- ❌ Parallel deployments/upgrades (too many variables)
- ❌ Pushing without user approval
- ❌ Syncing without user approval
- ❌ Rushing validation
- ❌ Skipping monitoring setup