Monitoring for unavailable segments taking ingestion tasks into account

We’d like to be able to alert when segment/unavailable/count is nonzero. But it looks like it spikes briefly whenever Kafka Ingestion Service tasks publish new segments. This makes sense, I think: this stat is only paying attention to historicals and not to KIS tasks.

I’d love to be able to alert a variant of this stat that only counts segments as unavailable if KIS isn’t also serving queries for them.

(a) Does something like this exist somewhere we haven’t found?

(b) How easy would it be to add it? Does the coordinator even track which segments are being served by KIS tasks somewhere, or is it a big architectural change?

–dave