[kwlug-disc] Prometheus for multiple server monitoring?

[John Sellens]

I'm a stick in the mud, but I still use nagios.

Coincidentally, I was musing today, just a few hours ago, that nagios
core hasn't changed that much in over 20 years.  I think Ethan's work
has held up pretty well.

I like that the configs are all in git, and can be automated if one is
inspired/dedicated enough.

Yes, the web interface is "traditional".

"Everyone" says "nagios sucks, prometheus rocks".  I look periodically,
and always find prometheus to appear to be so complicated, so many parts
to it.  Maybe I'm just too used to my own particular hell.

Lots of people seem to like zabbix, which I haven't looked at at all in
a long long time.

I also get the impression that prometheus is focused on metrics and
history, rather than events/alerts (up/down/slow).  In the old days,
nagios for events, and mrtg/cacti/cricket for graphs (for which I mostly
use collectd and friends recently).

I think about switching every once in a while, and then think that it
seems like a lot of work.


Cheers
John