How to manage the Prometheus service with systemctl

Managing the Prometheus service with systemctl keeps the monitoring server tied to normal Linux boot, maintenance, reload, and status workflows. A managed unit can start at boot, stop cleanly for planned work, reload scrape and rule changes, and report whether the web server is ready for queries.

A systemd unit named prometheus is the common target on Linux hosts that run Prometheus from a package. Ubuntu and Debian packages ship that unit with an /etc/default/prometheus environment file, an ExecStart command for /usr/bin/prometheus, and an ExecReload action that sends SIGHUP.

Use reload after validating configuration or rule-file changes, and use restart for startup flags, web configuration, retention arguments, or unit overrides. Check the /-/ready endpoint after a start or restart because an active process state does not always mean Prometheus has finished loading its TSDB and configuration.

Steps to manage the Prometheus service with systemctl:

Confirm that systemd knows the prometheus unit and read its reload action.

$ systemctl cat prometheus
# /usr/lib/systemd/system/prometheus.service
[Unit]
Description=Monitoring system and time series database
Documentation=https://prometheus.io/docs/introduction/overview/ man:prometheus(1)
After=time-sync.target

[Service]
Restart=on-abnormal
User=prometheus
EnvironmentFile=/etc/default/prometheus
ExecStart=/usr/bin/prometheus $ARGS
ExecReload=/bin/kill -HUP $MAINPID
TimeoutStopSec=20s
SendSIGKILL=no
##### snipped #####
[Install]
WantedBy=multi-user.target

The ExecReload line means systemctl reload prometheus sends SIGHUP to the running server. Custom units without ExecReload need a unit update or a restart path instead.

Enable Prometheus at boot and start it now.

$ sudo systemctl enable --now prometheus
Synchronizing state of prometheus.service with SysV service script with /usr/lib/systemd/systemd-sysv-install.
Executing: /usr/lib/systemd/systemd-sysv-install enable prometheus
Created symlink '/etc/systemd/system/multi-user.target.wants/prometheus.service' -> '/usr/lib/systemd/system/prometheus.service'.

If the unit is already enabled, systemd may print less output. The --now flag still asks systemd to start the service immediately.

Confirm the boot-enabled state.

$ systemctl is-enabled prometheus
enabled

Confirm the running process state.
```
$ systemctl is-active prometheus
active
```
If the state is failed or inactive after a start attempt, inspect the journal before checking the web endpoint.
Check the local readiness endpoint.
```
$ curl --silent --show-error http://127.0.0.1:9090/-/ready
Prometheus Server is Ready.
```
Use the host and port from the service's --web.listen-address flag when Prometheus does not listen on the default local port.
Reload Prometheus after a validated configuration or rule-file change.
```
$ sudo systemctl reload prometheus
```
Prometheus reloads configuration and rule files when it receives SIGHUP. Validate YAML with promtool before reloading a changed file.

Related: How to reload Prometheus configuration
Related: How to test Prometheus configuration
Confirm the service stayed active after reload.
```
$ systemctl is-active prometheus
active
```
Restart Prometheus when startup flags, storage retention, web configuration, or unit overrides changed.
```
$ sudo systemctl restart prometheus
```
A restart stops this Prometheus instance briefly. Use a redundant pair or a planned maintenance window when query or scrape continuity matters.

Related: How to set Prometheus retention

Check detailed service status after the restart.

$ systemctl status prometheus --no-pager
● prometheus.service - Monitoring system and time series database
     Loaded: loaded (/usr/lib/systemd/system/prometheus.service; enabled; preset: enabled)
     Active: active (running) since Sat 2026-06-20 10:32:58 UTC; 2s ago
       Docs: https://prometheus.io/docs/introduction/overview/
             man:prometheus(1)
   Main PID: 378 (prometheus)
      Tasks: 12
     Memory: 46.2M
        CPU: 92ms
     CGroup: /system.slice/prometheus.service
             └─378 /usr/bin/prometheus

Jun 20 10:32:58 linux-host prometheus[378]: ts=2026-06-20T10:32:58.263Z caller=main.go:1345 level=info msg="Loading configuration file" filename=/etc/prometheus/prometheus.yml
Jun 20 10:32:58 linux-host prometheus[378]: ts=2026-06-20T10:32:58.263Z caller=main.go:1384 level=info msg="Completed loading of configuration file" filename=/etc/prometheus/prometheus.yml
Jun 20 10:32:58 linux-host prometheus[378]: ts=2026-06-20T10:32:58.264Z caller=main.go:1124 level=info msg="Server is ready to receive web requests."

Inspect the service journal when start, reload, or restart does not reach the expected state.

$ sudo journalctl -u prometheus --no-pager
Jun 20 10:32:04 linux-host systemd[1]: Started prometheus.service - Monitoring system and time series database.
Jun 20 10:32:04 linux-host prometheus[77]: ts=2026-06-20T10:32:04.044Z caller=main.go:1124 level=info msg="Server is ready to receive web requests."
Jun 20 10:32:19 linux-host systemd[1]: Reloading prometheus.service - Monitoring system and time series database...
Jun 20 10:32:19 linux-host prometheus[77]: ts=2026-06-20T10:32:19.875Z caller=main.go:1345 level=info msg="Loading configuration file" filename=/etc/prometheus/prometheus.yml
Jun 20 10:32:19 linux-host systemd[1]: Reloaded prometheus.service - Monitoring system and time series database.
##### snipped #####

Parser errors, bad startup flags, permission errors, and TSDB startup messages usually appear in the unit journal before they appear in the web UI.

Stop Prometheus only when the instance should stop scraping targets and answering queries.
```
$ sudo systemctl stop prometheus
```
Stopping the service leaves dashboards, alert evaluations, and remote-write traffic without this Prometheus instance until it starts again.

Confirm the stopped state.

$ systemctl is-active prometheus
inactive