Seems like we have Watchdog alert group called example-group, might be better to call it something else like core-group maybe?
https://github.com/prometheus-operator/kube-prometheus/blob/44008b2682546fa738af5967ddfad86cd69b3e37/examples/existingrule.yaml#L1-L4
Lets agree on naming and would be great first issue I think.
cc @paulfantom @brancz
Probably not that relevant, but what do you think of changing the alertname to "DeadManSwitch"? I feel like that's the standard for this kind of alert, but I don't have strong opinions on keeping/changing the alertname
It was originally that, but we switched it to Watchdog to be gender neutral. Watchdog has equivalent etymology: https://en.wikipedia.org/wiki/Watchdog_timer
I agree it shouldn't be example-group. I'm happy with anything other than that really, core-group or literally watchdog works for me as well as I don't imagine anything else ever being in that group and if, then we can still change it.
We shouldn't have this alert in this place at all. This file is only used to demonstrate an example from https://github.com/prometheus-operator/kube-prometheus/blob/44008b2682546fa738af5967ddfad86cd69b3e37/docs/developing-prometheus-rules-and-grafana-dashboards.md#pre-rendered-rules.
The Watchdog alert we are shipping is put in general.rules group as seen in https://github.com/prometheus-operator/kube-prometheus/blob/main/manifests/kube-prometheus-prometheusRule.yaml#L14-L35
My proposal is to reinforce the fact that a file is only an example and change content to sth like:
groups:
- name: example-group
rules:
- alert: ExampleAlert
expr: example_sli < 99.999
labels:
severity: "critical"
My proposal is to reinforce the fact that a file is only an example and change content to sth like:
That sounds good to me as well! I would say we just use vector here so it does work out of the box:
groups:
- name: example-group
rules:
- alert: ExampleAlert
expr: vector(1)
labels:
severity: "warning"
SGTM, anyone would like to create a PR? :)
Was hoping this could be a good first issue for users as its not urgent fix.
Happy to take over this issue :slightly_smiling_face: