Skip to content

Kubernetes

Kubernetes is the control plane for running, scaling, and updating containerized workloads reliably.

This section focuses on practical operations: how resources behave, how rollouts happen, and how to debug failures quickly.

Scope

The Kubernetes layer covers:

  • Core workload resources and lifecycle
  • Declarative updates and rollout safety
  • Service discovery and networking basics
  • Storage and configuration patterns
  • High-signal debugging workflows

Mental Model

Most Kubernetes incidents can be traced to one of these categories:

  1. Desired state is wrong (manifest/config issue)
  2. Scheduler cannot place Pods (resources, constraints, taints)
  3. Runtime cannot start containers (image, command, permissions)
  4. Network path is broken (Service, DNS, policy)
  5. Readiness/health gates block traffic
  6. Dependencies are unavailable (database, API, storage)

Sections

Deployments

Understand:

  • How rolling updates and rollbacks work
  • How ReplicaSets and Pods map to rollout state
  • How to scale and restart safely
  • How to diagnose stalled rollouts quickly

DaemonSets

Understand:

  • How to run one Pod per eligible node
  • How update strategy affects node-agent rollouts
  • How to verify node coverage and placement
  • How to debug scheduling and rollout failures quickly

Authentication and Authorization

Understand:

  • How API requests are authenticated and authorized
  • How to validate effective permissions with kubectl auth can-i
  • How to apply least-privilege RBAC with Roles and Bindings
  • How to troubleshoot 401 Unauthorized vs 403 Forbidden quickly

Guiding Principle

When debugging Kubernetes:

  • Start with desired vs actual state.
  • Follow the object chain: Deployment -> ReplicaSet -> Pod -> Container.
  • Use events and conditions before assumptions.

This keeps diagnosis fast and repeatable.