Important

You are viewing documentation for an older version of Confluent Platform. For the latest, click here.

Troubleshooting alerts

Issue: Email alerts are prevented from being sent.

If sending=prevented appears in the response and logs, the email alerts feature needs to be enabled in the properties file. The feature is not enabled by default.

sending=prevented email to={} with subject={} from trigger={} messageId=

[2019-05-15 14:25:10,160] DEBUG [control-center-heartbeat-0] sending=prevented email to=recipient@email.com with subject

Resolution: Ensure that the confluent.controlcenter.email.enabled is set to true in the applicable control-center.properties file.

See Enable email alerts.

Issue: Webhook notifications are not being sent.

Resolution: Ensure that the confluent.controlcenter.webhook.enabled is set to true in the applicable control-center.properties file. The option is enabled by default.

Issue: The alert body (email) shows an unreachable host for the Control Center URL to alerts history.

Resolution: Ensure that the confluent.controlcenter.rest.listeners property is set. The first entry becomes the host:port combination in the alert body. See HTTPS Settings.

Note

The alerts feature checks for the presence of the confluent.controlcenter.rest.advertised.url configuration first. If that is empty, then it checks confluent.controlcenter.rest.listeners. If that is also empty, then the default value localhost:9021 is used. As a recommended best practice, configure both of these options. See Control Center Configuration Reference.

Issue: The cluster down action for the Control Center cluster was not created.

This is evident from a DEBUG log entry Failed to create cluster down action.

Resolution: Ensure that the cluster down configuration properties are properly configured. See Control Center cluster down status for detailed instructions for each type of notification action.

Issue: The cluster down alert for the Control Center cluster did not appear.

Reason: Depending on the timing of how quickly the cluster for Control Center went down, the alert might not have had time for any entry to be logged or for its trigger to be fired and delivered through its configured actions.