Operate
- Start from workload SLOs, failure model, team ownership, and platform constraints.
- Treat Kubernetes as a distributed control plane, not merely a YAML bucket.
- Prefer operational clarity over abstract platform cleverness.
- Design for rollout safety, debuggability, and failure containment.
Default Standards
- Requests/limits should be intentional.
- Health probes must reflect real readiness.
- Rollouts need safe defaults and rollback posture.
- RBAC and network policy should be least-privilege by default.
- Observability and runbooks are part of workload design.
References
- Workload design: references/workload-design.md
- Pods, deployments, and rollout safety: references/pods-deployments-and-rollouts.md
- Services and ingress: references/services-and-ingress.md
- Networking and network policy: references/networking-and-network-policy.md
- Storage and stateful workloads: references/storage-and-stateful-workloads.md
- Security and RBAC: references/security-and-rbac.md
- Resource management: references/resource-management.md
- Scheduling and autoscaling: references/scheduling-and-autoscaling.md
- Observability: references/observability.md
- Debugging and troubleshooting: references/debugging-and-troubleshooting.md
- Multi-tenant clusters: references/multi-tenant-clusters.md
- Platform operations: references/platform-operations.md
- Disaster recovery: references/disaster-recovery.md
- GitOps and delivery workflows: references/gitops-and-delivery-workflows.md
- Incident runbooks: references/incident-runbooks.md