Services
Start, stop, restart and reload the units installed on this server, and decide which come back after a reboot.
Services is a systemd front end for the units OSS Panel installed — the web server, the database engines, the cache, the process supervisor, PHP-FPM and fail2ban. It answers two questions at a glance: is everything running, and what is it costing in memory.
The table
| Column | What it tells you |
|---|---|
| Service | The friendly name, with the systemd unit underneath (nginx, php8.4-fpm). |
| Status | Running, Stopped or Failed. |
| Memory | Resident memory the unit is using right now. |
| CPU | Its share of the processor at the moment of the check. |
| Start on boot | Whether systemd will start it after a reboot. |
| Actions | Restart, plus a … menu. |
The banner at the top summarises the column — All 8 services are running — with the time of the last check and a refresh button next to it. The figures are a snapshot, not a live graph; refresh to re-read them.
Starting, stopping and restarting
Restart is on every row because it is the thing you need most often. The … menu holds the rest — Start, Stop, Reload, and View logs.
View logs jumps to that unit in System logs, which is usually where you are heading anyway when a service has stopped.
Anything disruptive confirms first, and the confirmation says what breaks:
- Stop — "Anything that depends on this service will stop working until you start it again."
- Restart — "It stops and starts again. Anything connected to it is cut off for a moment."
- Stop starting on boot — "It keeps running now, but it will not come back after the next reboot."
Start on boot
Most rows have a switch. Some show a padlock and Always on instead — those are the units the panel itself depends on. Nginx, Redis and PHP-FPM are in that group on a standard install: the panel runs on them, so it will not let you take them out of the boot sequence or stop them from here.
Reload where you can, restart where you must
Reload asks the service to re-read its configuration without dropping connections — the right choice for Nginx after a config change. Restart cuts everything connected to it for a moment. The panel offers reload only on units that actually support it.
Limits and caveats
- This lists the units the panel knows about, not every unit on the box. A
service you installed yourself over SSH will not appear here; manage it with
systemctl. - A Failed status is systemd's verdict, not a diagnosis. Open the unit's logs to find out why.
- Memory and CPU are point-in-time. For trends over a day, use the dashboard or, for database engines, Database health.
- Installing a new engine or PHP version adds its unit to this table automatically — see Database engines and PHP.