Kubernetes production checklist

Why a checklist?

After years working on Kubernetes clusters in high-concurrency environments, I have learned that the difference between a smooth deploy and a disastrous one often comes down to details forgotten in the rush.

The basics

• Separate namespaces for each environment (dev, staging, prod)
• Resource limits defined for every container
• Liveness and Readiness probes properly configured
• PodDisruptionBudget for critical services

Security

• Network policies active
• RBAC configured with least privilege principle
• Secrets managed with a secret manager (not in plain text in the repo)
• Image scanning enabled in the registry

Observability

• Centralised logging (ELK or equivalent)
• Metrics exposed and collected
• Alerting configured for critical KPIs

Conclusion

This list is not exhaustive, but it covers the points that have cost me the most time when I neglected them.