griesbacher / check_prometheus Goto Github PK

Hi,
can you pls elabore a bit how "health" check should work?
I have 19 targets down from 19 and check is still ok.

This is the result: "OK - There are 0 healthy and 19 unhealthy targets".

thx
Jan

Hi,

i'm just a ops guy which doesn't have the knowledge about go. Do you can documentate, how to build this project?

And if so, it possible to add precompiled binaries for coming releases?

Hi,
At the following line:

you use the warn value for Warning and Critical. That looks wrong.
If I use the check I get all the time a critical state and no warning state.
Regards,
René

Hi,

Unable to use search and replace when query contains rate, as metric name does not get included in perfdata output section.

eg;-
###works fine without rate name present in perfdata
check_prometheus m q --address http://localhost:9092/stage/prometheus -q 'node_vmstat_pswpin{instance="localhost:9100",job="core_node_exporter"}' -w '1:'
OK - Query: 'node_vmstat_pswpin{instance="localhost:9100",job="core_node_exporter"}'|'{name="node_vmstat_pswpin", instance="localhost:9100", job="core_node_exporter"}'=11660;1:;;;

###with rate specified in query name no longer present in perfdata
~/lib/monitoring-plugins$ check_prometheus m q --address http://localhost:9092/stage/prometheus -q 'rate(node_vmstat_pswpin{instance="localhost:9100",job="core_node_exporter"}[5m])' -w '1:'
Warning - Query: 'rate(node_vmstat_pswpin{instance="localhost:9100",job="core_node_exporter"}[5m])'|'{instance="localhost:9100", job="core_node_exporter"}'=0;1:;;;

Is there a way round this?

Thanks

Mark