What if sensor restarts?

The Sensor is designed to handle restart scenarios and will continue functioning normally after restart. No customer intervention is needed until the sensor crashes. Sensor will restart according to the sensor restart policy.

Sensor restart policy

Exceptions will be handled gracefully and the sensor will restart as per its restart policy for recoverable and irrecoverable errors, as described below.

Recoverable errors

The sensor will return a recoverable error code 24 in cases like the sensor has crashed or the sensor caught an exception. In these cases, the sensor will recover on its own, and will keep on restarting. There is no max limit set on the number of restarts, but the time between two restarts will increase with the number of restarts needed, up to 16 minutes. For example, the time between two restarts could be 1 minute to start, then 2 minutes, then 4 minutes, then 8 minutes, then 16 minutes. Once 16 minutes is reached, the time between restarts will remain at 16 minutes. No core dump file will be created.

Irrecoverable errors

If the sensor returns an irrecoverable error code, it means the sensor will not recover on its own and the sensor will exit. For standalone deployments, the sensor will exit upon receiving the irrecoverable error code. For DaemonSet deployments, when the sensor exits with an irrecoverable error code, the Kubernetes Pod restart policy will restart the exited container. Irrecoverable error codes must be resolved by making changes to the deployment files and deployment arguments.