Changes between Version 1 and Version 2 of Ticket #29864, comment 3


Ignore:
Timestamp:
Jun 13, 2019, 8:10:34 PM (4 months ago)
Author:
anarcat
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #29864, comment 3

    v1 v2  
    88a key takeaway from the presentation is that they didn't replace nagios, but filtered its output. this matches our model where nagios sends notifications to alert manager.
    99
    10 we might also want to monitor the number of alerts sent out by nagios, if only to figure out trends that we could work to fix.
     10we might also want to monitor the number of alerts sent out by nagios, if only to figure out trends that we could work to fix. for this I found two different solutions:
     11
     12 1. a [https://github.com/m-lab/prometheus-nagios-exporter simple exporter] relying on the [https://checkmk.com/cms_livestatus.html check-mk-livestatus] which [https://tracker.debian.org/check-mk-livestatus did not make it to buster]
     13 2. a [https://github.com/Griesbacher/Iapetos more complex exporter] that behaves like a broker but that is not packaged in Debian either
     14
     15It seems neither is a good solution for now and that we should instead rely on the internal reporting metrics inside Nagios for the time being.