Having many new alerts can come as a burden for SRE and cluster admins. They might be interested in an alternative way of getting the Health information.

The configured alerts could have a "switch" in the FlowCollector CRD to generate Recording Rules instead. When this switch is enabled, no alert is created. The exact same promQL query is then performed, but as part of a recording rule.

The console plugin needs then to fetch the recording rule definitions, and run them, to get metrics values. In that case, there is no such concept of "firing", "pending", "silenced".