fix: alertmanager metamon
What
Filter alertmanager notification failures only for ops
instance.
Why
We only have a single alertmanager cluster inside of the ops
environment.
In gitlab-com/gl-infra/production#14467 (closed) we had a bad deployment here we had alertmanagers in different clusters by mistake which lead to pages for the on-call and it was a FalseAlarm.
Scoping this just for the env="ops"
will help prevent incidents like
this from happening again.