The Services page, in the Server group, lists the system services that run the sites:
| Row | System service |
|---|---|
nginx | nginx |
mariadb | mariadb |
redis | redis-server, only where it is installed |
| Site PHP | every cloudground-php-*.service, one per site |
For each service you see State, At boot (yes or no), Memory, Started and Version. The Site PHP row sums up the PHP processes: how many are running and, when one is not, which.
A service that is not installed has no row. A new server does not get Redis (the sites' object cache
uses APCu), so there is no redis row. A server upgraded from an earlier version keeps Redis switched
off, and its row says disabled. The nginx and mariadb rows are always there.
Read the state
| State | Meaning |
|---|---|
| running | The service is running. |
| failed | The service stopped with an error. |
| disabled | The service is off and does not start on boot, on purpose. |
| stopped | The service is not running, but it is meant to. |
A failed or stopped service raises an alert on the dashboard: see alerts and notifications.
Restart a service
- Open Services.
- Press Restart on the service's row.
- Confirm.
A restart interrupts requests in flight for a few seconds. The panel waits for the system to finish, 3 minutes at most: MariaDB can be slow to come back.
A disabled service has no Restart button: it is off on purpose (Redis, on an upgraded server) and a restart would turn it on. The Site PHP row has no Restart either.
From the API
GET /api/services reads the list; POST /api/services/<name>/restart restarts nginx, mariadb
or redis. The restart of a service that is not installed is refused, and nothing is started, with
409 <name> is not installed on this server; that of a disabled service with 409 too. Every
restart lands in the Audit log.
Next step
When a service stops, the reason is in the logs.