Kubernetes Operators Best Practices

Kubernetes Operators are processes connecting to the master API and watching for events, typically on a limited number of resource types. When a relevant event occurs, the operator reacts and performs a specific action. This may be limited to interacting with the master API only, but will often involve performing some action on some other systems (this could be either in cluster or off cluster resources).

Operators are implemented as a collection of controllers where each controller watches a specific resource type. When a relevant event occurs on a watched resource a reconcile cycle is started. During the reconcile cycle, the controller has the responsibility to check that current state matches the desired state described by the watched resource.

Interestingly, by design, the event is not passed to the reconcile cycle, which is then forced to consider the whole state of the instance that was referenced by the event. This approach is referred to as level-based, as opposed to edge-based. Deriving from electronic circuit design, level-based triggering is the idea of receiving an event (an interrupt for example) and reacting to a state, while edge-based triggering is the idea of receiving an event and reacting to a state variation.

Source: openshift.com