Kubernetes Mastery (Senior → Principal)

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