@thomas I know none of them, but I'm CheckMK... let's say expert.
I don't really get what you're trying to achieve. You've got 3 instances of a service that should aggregate into a single metric? Like: Counting files on 3 different machines and receiving the metric of the sum of all files?
@FrankyFire the network connection on my monitoring host seems to be a bit flaky, esp. At night. I'm receiving a lot of false alarms for my ping checks, so I was hoping to get some decentralized setup running... With multiple servers to check from. Maybe they could then decide amongst themselves if a service is really up or down 🤔
@thomas this is possible and should be possible with Icinga too (as their based on the same technology). In CheckMK I would do the following:
Set up the ping check from multiple instances (also outside of your network) and create either a cluster of those hosts, or use Business Intelligence to create an aggregation with parameters like 'only crit of all unavailable'. Something similar should be possible in any monitoring solution.
@thomas More a workaround than a solution, but you could just set the regular ping checks from A and B to silent and add another check that checks the status of the other checks. if all of them != OK, return ICINGA_CRITICAL.
Metalhead.club is a Mastodon instance hosted in Germany and powered by 100% green energy.