The orchestrator's model of the world

the API that decides what should run went down, and everything already running carried on serving traffic perfectly.

Control plane

the nodes running your containers and the proxies moving the packets keep working whether or not anyone is watching them.

Data plane

a machine joins the pool, advertises how much CPU and memory it has, and starts being given other people's workloads.

Node

you stopped addressing individual machines and started submitting work to a pool that decides where it lands.

Cluster

you ran the commands in order, and doing it again would create a second copy of everything.

Imperative configuration

you write down what should exist, submit it, and something else works out the steps — including that most of them are already done.

Declarative configuration

you said three replicas, and every time someone deletes one a fourth appears without anybody being asked.

Desired state

the spec says three and the status says one, and the gap between those two numbers is the entire incident.

Observed state

something reads the difference between what you asked for and what exists, takes one step to close it, and then does it all again forever.

Reconciliation loop

you never create the individual containers; you create an object, and a background process creates them to match it.

Controller

the failover, the backup and the version upgrade for the database are performed by software encoding what the DBA used to do.

Operator

your team's own object type — a queue, a tenant, an environment — is created the same way as anything else and reconciled by your own controller.

Custom resource

you asked for the workload to run somewhere, and something scored every node and picked one before any container existed.

Scheduler

two containers share an IP address and a lifetime, and are scheduled onto the same node together or not at all.

Pod

a second container next to your application handles TLS, logging or metrics, and your code was never changed.

Sidecar

a throwaway container runs the migration, finishes, and only then does the application container start.

Init container

the log collector appears on every node automatically, including the one that joined the cluster four minutes ago.

DaemonSet

the replicas are named zero, one and two rather than randomly, and each one gets its own disk back after a restart.

StatefulSet

the pod died, came back on another node, and the data was still there because the disk followed it.

Persistent volume