I have uploaded a new version of Plack::Middleware::Statsd to #CPAN https://metacpan.org/release/RRWO/Plack-Middleware-Statsd-v0.6.3

This is a #Perl #Plack plugin for #statsd monitoring.

The new version improves some of the documentation and reverts back to using Try::Tiny for error trapping.

I've no idea if any of this is "correct," but I've got a New Relic dashboard that shows Postgres table stats for this Mastodon server. Currently it's got the row count and table size of the Accounts and Statuses tables.

Python source for the data collection is here: http://i.abackstrom.com/share/2023/pgstats.py.txt

#MastoAdmin #NewRelic #StatsD

For those interested: I'm using #Prometheus (https://prometheus.io/) and the #redis, #postgres, and node exporter to monitor various components of Mastodon and the servers I run, along with the #statsd exporter to gather Mastodon-specific metrics.

I'm also using the wonderful dashboards from @IPngNetworks with #grafana to visualize our instance's health; strongly recommend checking those out as well here:
https://ipng.ch/s/articles/2022/11/27/mastodon-3.html

I have updated Plack::Middleware::Statsd https://metacpan.org/release/RRWO/Plack-Middleware-Statsd-v0.6.2

This is #Perl #Plack middleware for logging metrics to #statsd.

This change includes some minor code improvements, and updates to the documentation.

I am so utterly confused at how #prometheus, #grafana (and especially Grafana cloud), and #statsd interact. I THINK Prometheus polls... something? And then writes data somewhere, and somehow Grafana displays it?

That implies that my service writes metrics to a listener service, which then runs an endpoint of its own so yet another service can ingest the metric results, which are then written to a storage system somewhere in the cloud so that Grafana can display them.

That can't be right.

And I got it all running. Never used #grafana, #statsd or #prometheus before. Turned out to be fairly simple

I built a small project to help with debugging/logging StatsD metrics on a local enviroment (really just dumping them to STDOUT) - find it on GitHub and Docker Hub:

https://github.com/jeanreis/statsd-dumper
https://hub.docker.com/repository/docker/jeanreis/statsd-dumper

#statsd #metrics #docker

@rbairwell I recently noticed that mastodon does support #statsd statistics, so you could send that to #influxDB and dashboard it in #grafana. Make it public, and carefully make some dashboards

#mastodon #statsd #prometheus #servicemonitor #kubernetes

Is anyone using the #statsd metrics reporting functionality of #mastodon? I just enabled it and I'm making graphs in grafana but I'm not sure I understand all the metrics that are being reported. #mastoadmin

Christian Clauss is working on moving pytatsd to Github actions for automated testing and builds https://github.com/sivy/pystatsd/pull/109 #python #statsd #opensource

(monkinetic txt-5ae8560)

@kelbot

I'm a big fan of #statsD, which is an etsy #foss project. That might be a bit different from what you're asking about, but I thought I'd mention it (this being #fosstodon and all!)