mirror of
https://github.com/fleetdm/fleet
synced 2026-04-21 13:37:30 +00:00
f761827850
Move CalculateAggregatedPerfStatsPercentiles reads to the replica #18838 I manually tested the aggregated query stats change by using a read replica. https://github.com/fleetdm/fleet/blob/main/tools/mysql-replica-testing/README.md - But set `master_delay=0` due to issue https://github.com/fleetdm/fleet/issues/19272 - Run a saved query as a live query, and see that its stats in `aggregated_stats` table are updated. # Checklist for submitter - [x] Changes file added for user-visible changes in `changes/`, `orbit/changes/` or `ee/fleetd-chrome/changes`. See [Changes files](https://fleetdm.com/docs/contributing/committing-changes#changes-files) for more information. - [x] Added/updated tests - [x] Manual QA for all new/changed functionality
5 lines
588 B
Text
5 lines
588 B
Text
MySQL query optimizations:
|
|
- During software ingestion, switched to updating last_opened_at as a batch (for 1 host).
|
|
- Removed DELETE FROM software statement that ran for every host update (when software was deleted). The cleanup of unused software is now only done during the vulnerability job.
|
|
- `/api/v1/fleet/software/versions/:id` endpoint can return software even if it has been recently deleted from all hosts. During hourly cleanup, this software item will be removed from the database.
|
|
- Moved aggregated query stats calculations to read replica DB to reduce load on the master.
|