how to make prometheus as system service

[Amari CH]

Currently I just nohup run the processes in backend, this is ugly.
I am using ubuntu 18.04 system, how to make those processes to
start/stop as system daemons?

Thank you.

[sayf.eddi...@gmail.com]

this is my systemd file 

```

[Unit]

After=network.target

Description=prometheus server

Documentation=http://prometheus.io

[Service]

User=prometheus

Type=simple

EnvironmentFile=/etc/sysconfig/prometheus

ExecStart=/usr/bin/prometheus --config.file=/etc/prometheus/prometheus.yml $EXTRA_ARGS

ExecReload=/bin/kill -HUP $MAINPID

Restart=on-failure

RestartSec=10

SuccessExitStatus=0

LimitNOFILE=3000000

[Install]

WantedBy=multi-user.target

```

Notice this is for centos/rhel, by the standard the environment files are under `/etc/default/prometheus` for debian based OSs, the file should contain the rest of the config (tsdb path web address and port etc)

Drop this as /usr/lib/systemd/system/prometheus.service and run `systemctl enable prometheus`

[Ben Kochie]

Try https://github.com/cloudalchemy/ansible-prometheus

--
You received this message because you are subscribed to the Google Groups "Prometheus Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to prometheus-use...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/f8588cc5-8963-4ed0-e9a0-99d3e02ddcae%40maddoghost.com.

[Brian Candler]

Under [Service] I'd also recommend

TimeoutStopSec=300

This is to give prometheus more time to finish writing out its WAL if you do "systemctl stop" or "systemctl restart".

[sayf.eddi...@gmail.com]

Nice tip, thanks !

[Amari CH]

Is there an easy way to migrate prometheus into apt sources so that we
can run apt update and apt install|upgrade to upgrade them?

Thanks

--
Amari CH
https://maddoghost.com/

[Brian Candler]

On Thursday, 17 September 2020 09:07:45 UTC+1, Amari CH wrote:

Is there an easy way to migrate prometheus into apt sources so that we
can run apt update and apt install|upgrade to upgrade them?

There are existing prometheus packages in the Ubuntu repositories.  Beware they may be ancient: e.g. Ubuntu 18.04 has prometheus-node-exporter 0.15.2, and prometheus 2.1.0.  Still, they may be adequate.

All you really need to deploy is the node_exporter binary and the systemd unit file to each host.  You can just copy those two files, either manually or using whatever configuration management tools you may already be using.

[Amari CH]

Brian Candler wrote:
>
> All you really need to deploy is the node_exporter binary and the
> systemd unit file to each host.  You can just copy those two files,
> either manually or using whatever configuration management tools you may
> already be using.

Brian, do you have a biz solution for hosting industry based on
prometheus? we can pay for it. thanks.

[Brian Candler]

No, but see https://prometheus.io/support-training/