reconciliation

The repeated check-and-fix loop in which a program compares desired state with observed state and changes the world to match.

In more detail

Every Kubernetes controller runs this loop: watch the objects it owns, notice a difference, take one step to reduce it, and check again. Because the loop keeps running, a deleted pod or a lost node is repaired without anyone typing a command. GitOps tools and operators use the same idea with a Git repository or a custom resource as the source of desired state.

Examples

A Kubernetes ReplicaSet controller sees two pods where three are wanted and creates one.
In Terraform: Each Terraform plan proposes a reconciliation between configuration and reality, and apply carries it out. The same idea drives Kubernetes controllers, which reconcile continuously instead of when someone runs a command.

Related terms

Taught in

References

Certification courses