Learn · Operations

Operations

Operations is a protocol. Under pressure you need sequences that preserve attribution: observe → narrow → act → confirm → memorialize.

What this module covers

Rites that survive incidents.

Evidence-first workflows: describe/events/logs, then targeted diffs and rollouts.
Change discipline: smallest-safe fixes, reversible rollouts, and explicit verification.
Containment: reducing blast radius while you regain signal.

Operator-grade and precise.

Chapter

Operations Handbook

Debugging pods, rollouts, logs, events, namespaces, and failure modes.

Text

Incident Doctrine for Platform Teams

Containment, communication, reversibility, and the discipline of truth.

Text

Debugging the Control Plane Under Pressure

API saturation, admission failures, controller churn—and how not to amplify the outage.

Exercises that build calm execution.

Atlas

Node NotReady

Containment, cordon/drain posture, and failure-domain thinking.

Atlas

OOMKilled and Evictions

Prove the killer, then right-size requests/limits with restraint.

Lab

Pending Pods (Scheduling Constraints)

Read scheduler events; fix the governing constraint; verify placement.

Continue with security and reliability posture.

Learn

Security & Reliability

RBAC, policy, secrets handling, observability posture, and stable trust boundaries.

Topic map

Security (Archive)

Curated doctrine and practice links for trust boundaries.