Not every failed job warrants an alert. The new Margins alerting feature tells Sentry when to mark a job as failed or missed, and Failure and Recovery Thresholds define when to create or resolve an issue.

Margins give you more control over what Sentry marks as a missed or failed check-in. You can define a Grace Period, a set number of minutes before a check-in is considered missed, and a Max Runtime for the number of minutes before an in-progress check-in is marked as timed out.

Failure Thresholds will create a new issue when a set number of missed or error check-ins are processed. Recovery Thresholds do the opposite - automatically closing out issues after a set number of consecutive healthy check-ins are processed.