Notification

Replicator has the ability to send notifications to configured backends when critical issues are detected such as when a worker pool or job group enters failsafe mode.

Supported Backends

Replicator employs the factory/provider model for notifiers, therefore if your required backend is not currently available please either raise a feature request or submit a PR for this work. The current supported backends are:

• PagerDuty - Configuring the PagerDuty notifier will allow Replicator to raise incidents when worker pool scaling or job group scaling enters failsafe mode. The CLI flag pagerduty-service-key is the only flag required to setup the PagerDuty notifier and the key represents an Events API v1 service integration key.

Notification Enhancements

When sending alerts and notifications it is important to be as clear and as descriptive as possible. Therefore Replicator has a number of configuration parameters to make notifications easy to understand:

• replicator_notification_uid - This parameter should be configured within both the worker pool meta parameters and the job group meta parameters and is designed to be a UID which represents a runbook entry for operators to follow.

• -cluster-identifier - This CLI flag and configuration parameter allows operators to define a human readable cluster identifier which easily and clearly defines which cluster the notification represents.