Chapter 135. AutoRestart schema reference

Used in: KafkaConnectorSpec, KafkaMirrorMaker2ConnectorSpec

Full list of AutoRestart schema properties

Configures automatic restarts for connectors and tasks that are in a FAILED state.

When enabled, a back-off algorithm applies the automatic restart to each failed connector and its tasks. An incremental back-off interval is calculated using the formula n * n + n where n represents the number of previous restarts. This interval is capped at a maximum of 60 minutes. Consequently, a restart occurs immediately, followed by restarts after 2, 6, 12, 20, 30, 42, 56 minutes, and then at 60-minute intervals. By default, Streams for Apache Kafka initiates restarts of the connector and its tasks indefinitely. However, you can use the maxRestarts property to set a maximum on the number of restarts. If maxRestarts is configured and the connector still fails even after the final restart attempt, you must then restart the connector manually.

For Kafka Connect connectors, use the autoRestart property of the KafkaConnector resource to enable automatic restarts of failed connectors and tasks.

apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaConnector
metadata:
  name: my-source-connector
spec:
  autoRestart:
    enabled: true

If you prefer, you can also set a maximum limit on the number of restarts.

apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaConnector
metadata:
  name: my-source-connector
spec:
  autoRestart:
    enabled: true
    maxRestarts: 10

For MirrorMaker 2, use the autoRestart property of connectors in the KafkaMirrorMaker2 resource to enable automatic restarts of failed connectors and tasks.

apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaMirrorMaker2
metadata:
  name: my-mm2-cluster
spec:
  mirrors:
  - sourceConnector:
      autoRestart:
        enabled: true
      # ...
    heartbeatConnector:
      autoRestart:
        enabled: true
      # ...
    checkpointConnector:
      autoRestart:
        enabled: true
      # ...