collectd vs node-exporter

[Matt Zagrabelny]

Greetings,

I'm a brand new Prometheus user. Please excuse the question if it is foolish.

How much of a difference is there between collectd (and its exporter) and the node-exporter?

Do folks use both on their systems or do both of those exporters mostly overlap?

Thanks for the help!

-m

[Julius Volz]

The collectd exporter is meant more for supporting legacy setups that already have collectd. Most people just use the node exporter (and other exporters for more specific stuff).

--
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/c37cfc98-0733-458d-a937-d6730e033652%40googlegroups.com.

[Harald Koch]

I use node_exporter (and other prometheus exporters) for everything I can, but I'm still using collectd for two metrics:

* hard drive temperatures

* APC UPS metrics

--

Harald

[Brian Candler]

FWIW, there's a node_exporter textfile collector script for SMART stats including drive temperature:

https://github.com/prometheus-community/node-exporter-textfile-collector-scripts/blob/master/smartmon.sh

Also there are a couple of different APC UPS exporters which talk to apcupsd.  I have tried damomurf/apcupsd-exporter and mdlayher/apcupsd_exporter.  The former is the one I'm still using.

[Harald Koch]

So I had looked at the apcupsd exporters earlier, but didn't have much luck, but today they're working perfectly for me. C'est la vie, I guess :)

I have that old greybeard UNIX guy aversion to firing off a process each time I want to monitor something (which is one of the reasons why I'm replacing both Nagios/Icinga2 and Cacti with Prometheus) but I'll give that textfile collector script a try anyway...

Thanks so much!

--

Harald