Restarting a service can help troubleshoot deployment issues by restarting all containers and processes within the service. If any container images have been updated, a new image may be pulled, depending on the image pull policy set for the service.
After a service is restarted, it will follow the configured restart policy (see the service specification guide for restart policy options). If a container crashes (exits unexpectedly), it enters a crash loop backoff state, where a restart is attempted every 10 seconds for a maximum of 5 minutes (300 seconds). If the container crashes again, the backoff timer resets and attempts restarts at 5-minute intervals only (unless the service spec was changed). To exit the backoff state, the container must run for at least 10 minutes without crashing.
To restart a service:
- In the Secure Edge Portal, select Nodes > All Nodes, and then select the node to open its details page.
- On the Services tab, find the service you want to restart, click on the three-dot menu, and select the Restart Service icon.
- In the confirmation dialog, select Yes - Restart Service.
- The service status shows as RESTARTING. Once the restart is successful, the service status updates to HEALTHY.