Module 5
Workloads
You almost never create Pods directly. Deployments, StatefulSets, DaemonSets, Jobs and CronJobs are controllers that create them with different guarantees.
Why it matters. Choosing the wrong workload type causes specific, predictable failures: databases that lose identity, migrations that restart forever, agents that run on only some nodes.
After this module you can answer
- Why does a Deployment own ReplicaSets, not Pods?
- How does a rolling update stay available?
- What does a StatefulSet guarantee that a Deployment does not?
- When do I use a DaemonSet or a Job?
- What are init containers and native sidecars for?
What is inside
1Deployment and ReplicaSet
2Rolling updates
3StatefulSet
4DaemonSet, Job, CronJob
5Init and sidecar containers
6Choosing the right one
7Module infographic