datawave
datawave copied to clipboard
Update ActiveQueryLog to periodically send a list of running query scans from the tserver to another service
ActiveQueryLog keeps track of what queries have scans running on a tserver at any given time. This information could be useful from an operational perspective in order to determine if there are any orphaned scans (i.e. scans that are continuing to run even though the query is dead from a webservice perspective). This information could be sent off to another service via a fire-and-forget rest call and aggregated in something like hazelcast (which is commonly used by our microservices). We would then want to implement some kind of UI that correlates the active queries with the active scans. We could use this to determine which queries have the most scans running, which queries have the longest scans running, and which scans appear to be orphaned. Grouping and/or filtering by tserver instance would be useful as well. This would be a huge time saver during troubleshooting.