Different job identifiers are used between k8s and VMs for metrics
@andrewn comments on slack:
Just noticed this in passing: we’re using different job identifiers in Prometheus in k8s, which means that our monitoring won’t pick these metrics up: {job="gitlab-webservice",type="git"}
We can switch to using a regular expression match too, but this is definitely something to be aware of
we should make sure the job identifiers are aligned, especially for services like git
that are not yet in production, but will soon be migrated and be taking prod traffic in both k8s and vms.
MRs:
Edited by John Jarvis