//
Search across all documentation pages
9 pages in this section.
How the control-plane pieces fit together - the orientation before diving into each one
The 30,000-ft loop: you declare desired state; controllers make reality match it
Level-triggered control, desired vs observed state, and why controllers are eventually consistent rather than transactional
How a `kubectl apply` becomes running pods: API server → etcd → scheduler → controller-manager → kubelet
The single source of truth, quorum/consensus, and why etcd health *is* cluster health
Optimistic concurrency, owner references, and finalizers explained
Design apps and controllers for eventual consistency and retries, not instant reconciliation