Chris's Wiki :: blog/sysadmin/PrometheusScriptMetricsHow Commentshttps://utcc.utoronto.ca/~cks/space/blog/sysadmin/PrometheusScriptMetricsHow?atomcommentsDWiki2020-01-05T20:15:47ZRecent comments in Chris's Wiki :: blog/sysadmin/PrometheusScriptMetricsHow.By Chris Siebenmann on /blog/sysadmin/PrometheusScriptMetricsHowtag:CSpace:blog/sysadmin/PrometheusScriptMetricsHow:f3a7cbd67cba6778f8f1ee0b05df612dac80d367Chris Siebenmann<div class="wikitext"><p>I don't think very many (or any) of the metrics we're gathering through
scripts could be gathered through Telegraf. A lot of them are quite
custom local metrics that require things like parsing the output of
status reporting programs for Exim, SLURM, and OpenBSD's various VPNs.
I believe that all of them are pretty lightweight, and when they aren't
the most expensive portion is the external programs they have to run to
get the information.</p>
<p>(Some of the 'scripts' are compiled programs themselves, written in
Go.)</p>
</div>2020-01-05T20:15:47ZBy gmuslera on /blog/sysadmin/PrometheusScriptMetricsHowtag:CSpace:blog/sysadmin/PrometheusScriptMetricsHow:c63ba7ca7eb1e0d2063cb4cdff9289fe0e7441a9gmuslera<div class="wikitext"><p>Have you tought on using InfluxData's Telegraf as Prometheus exporter? Or just pushing the data when its collected to any of the supported databases, that can be queried by Grafana?</p>
<p>It may be less expensive in resources than launching a cron job and could collect a lot of different kinds of local and remote metrics.</p>
</div>2020-01-05T20:01:09Z