Hey folks! The druid docs suggest using the sys tables instead of using loadstatus API as mentioned here : https://druid.apache.org/docs/latest/api-reference/legacy-metadata-api/ But it’s unclear if the system tables provide cached or actual results. We need uncached results to prevent metric-level table ingestion failures. Would the querying of sys tables be cached or will that return the accurate representation of the system? We are currently using loadstatus and its causing heavy load on the coordinator causing a restart. Any help here would be appreciated.

The SYS schema is an in-memory schema populated by the services (Coordinator & Overlord?) which poll data periodically from other services and the RDS. I don't know the details of the data flow and polling sequences though, and I wouldn't think there would be addiional caching on top of the fact that this data is cached in memory to begin with ... but you could try setting up a separate connection to the cluster from a differenet web console in a different browser to see if you can spot any different in data between the two that would suggest some sort of caching behavior ...