I like the idea of Prometheus Operator where alerting configurations are deployable kubernetes objects defined with CRDs.
apiVersion: monitoring.coreos.com/v1
kind: PrometheusRule
metadata:
labels:
prometheus: example
role: alert-rules
name: prometheus-example-rules
spec:
groups:
- name: ./example.rules
rules:
- alert: ExampleAlert
expr: vector(1)
I'd like to do the same with Grafana Loki.
Loki`s docs mention only that its alerting rules are "compatible" with Prometheus. Yet, we still have to declare all the rules within single Loki configuration file. All the alerts should be published to Prometheus Alertmanager.
The Internet mentions about Loki operator, which in fact looks familiar to Prometheus Operator, but docs seems to be lacking a lot of details.
Does anyone managed to mimic prometheus operator behaviour with Loki as alert source?
