Tune new nginx prometheus alerts
We have new nginx metrics together with a few new prometheus alerts for them (!1166 (merged)).
But now we often get alerts for these two alerts:
- 4xx errors
- 499 errors
And very rarely alerts about 5xx errors: Job #229992 failed for 4dee2630:
I propose to
- Remove the 4xx error alert, since it if firing too often (and could also be triggered by scanning a cluster for random endpoints)
- Remove the 499 and 5xx error alerts, but add them to our stackspin.net (or maybe to staging.stackspin.net)
What do you think ?
Edited by Varac