If newly created instances for replacing the unhealthy ones in a scale set continue to remain unhealthy even after repeatedly performing repair operations, then as a safety measure the platform updates the *serviceState* for automatic repairs to *Suspended*. You can resume the automatic repairs again by setting the value of *serviceState* for automatic repairs to *Running*. Detailed instructions are provided in the section on [viewing and updating the service state of automatic repairs policy](#viewing-and-updating-the-service-state-of-automatic-instance-repairs-policy) for your scale set.
0 commit comments