This set builds vocabulary for extending Kubernetes with custom controllers and operational automation.
0 / 5 completed
1 / 5
At standup, a dev describes a custom controller that encodes operational knowledge to manage a complex stateful application's full lifecycle in Kubernetes. What is this pattern called?
The Operator pattern extends Kubernetes with a custom controller that encodes human operational knowledge, such as how to back up, upgrade, or scale a specific stateful application, automating tasks that would otherwise require manual intervention. This goes beyond basic built-in resource management. It is commonly used for complex systems like databases running on Kubernetes.
2 / 5
During a design review, the team defines a new resource type representing their custom application's desired state, extending the Kubernetes API. What is this called?
A Custom Resource Definition extends the Kubernetes API with a new resource type, letting an application's desired state be expressed and managed the same declarative way as built-in resources like Deployments. An Operator's controller then watches and reconciles this custom resource. CRDs are the foundation that makes the Operator pattern possible.
3 / 5
In a code review, a dev explains that the Operator continuously compares the actual state of the cluster to the desired state and takes corrective action. What is this loop called?
The reconciliation loop is the core control-loop pattern where a controller continuously observes the actual cluster state, compares it to the desired state defined in a custom resource, and takes corrective action to close any gap. This continuous, self-healing process is what distinguishes Kubernetes-native automation from a one-off imperative script. It runs indefinitely as long as the Operator is deployed.
4 / 5
An incident report shows an Operator repeatedly attempted an invalid corrective action in a tight loop, consuming excessive cluster resources. What is this failure mode called?
A buggy reconciliation loop that keeps attempting an action that cannot succeed, without backoff or proper error handling, can spin rapidly and consume excessive cluster resources. Well-designed operators include backoff strategies and clear failure states to avoid this. This failure mode is a known operational risk specific to continuously running control loops.
5 / 5
During a PR review, a teammate asks how an Operator differs from a plain Kubernetes Deployment for running a stateful application. What is the distinction?
A plain Deployment handles generic pod scheduling and replica management, while an Operator layers in application-specific operational knowledge, like how to safely fail over a database primary, that a generic resource type has no way to express. This makes Operators suited to complex, stateful systems with nontrivial lifecycle requirements. Simpler stateless applications often don't need this extra complexity.