CodePark/ questions / Orchestration & Kubernetes basics
START PRACTICINGPRACTICE
Backend question · Docker › Orchestration & Kubernetes basics

What does a Kubernetes Deployment do, and how does a rolling update work?

A Deployment describes the desired state of a stateless app. It then keeps the cluster in that state. The desired state includes:

  • which image to run
  • how many replicas (copies) to run
  • the Pod template

The Deployment manages ReplicaSets. A ReplicaSet keeps the right number of Pods running and replaces any that die.

How a rolling update works:

  • You change the Pod template, for example to a new image tag.
  • The Deployment creates a new ReplicaSet.
  • It slowly scales the new ReplicaSet up and the old one down.

Controlling the pace:

  • maxSurge sets how many extra Pods may exist during the update.
  • maxUnavailable sets how many Pods may be missing.
  • Both are 25% by default.

New Pods get traffic only after their readiness probe passes. This is a check that the app is ready to serve. So a broken version stops the rollout instead of taking the service down.

kubectl rollout status follows the progress. kubectl rollout undo rolls back to the previous ReplicaSet.

  • A Deployment declares the number of replicas and the Pod template.
  • It manages ReplicaSets, which keep the Pods running.
  • A rolling update moves Pods from the old ReplicaSet to the new one.
  • maxSurge and maxUnavailable control the pace of the update.
  • Readiness probes decide when Pods get traffic, and kubectl rollout undo rolls back.