Chapter Ten
Cluster Operations
Running a cluster over time — installing with kubeadm, operating and backing up etcd, upgrades and node maintenance, multi-tenancy and quotas, and managing cluster cost.
A demo cluster and a production cluster differ mostly in operations: the unglamorous work of installing, backing up, upgrading, and paying for the thing.
This chapter is that work. Standing up a cluster with kubeadm, treating etcd as the database it is and backing it up, upgrading control plane and nodes without downtime, carving the cluster among tenants with quotas, and the cost levers that keep the bill honest.
Topics in This Chapter
Topic 49
Installing with kubeadm
Bootstrapping a conformant cluster by hand — what kubeadm does, what it leaves to you, and why doing it once is worth it.
Topic 50
etcd Operation and Backup
The key-value store that holds all cluster state. Quorum, performance, and the backup-and-restore procedure you must rehearse.
Topic 51
Upgrades and Node Maintenance
Version skew rules, draining and cordoning nodes, and the rolling upgrade path that avoids an outage.
Topic 52
Multi-tenancy and Quotas
Sharing a cluster safely — namespaces, ResourceQuotas, LimitRanges, and the line where soft multi-tenancy stops being enough.
Topic 53
Cluster Cost Management
Where the money goes, how to attribute it, and the levers — right-sizing, bin-packing, spot nodes — that cut it without cutting reliability.