Description
Making query threads wait on output queues when they fill up means the query engine will be slower when the audit machinery cannot keep up. To aid in diagnosing this, we propose to keep records of the query-side audit behavior: how many audit records were received, how many were filtered out, how many were successfully sent, and how many errored out (all in the last minute). This information would be available from the /admin/stats settings, and should be viewable somewhere in the admin console. (Proposed for a future version.)
See also: MB-28332
Attachments
Issue Links
- relates to
-
MB-28332 determine error rate of audit transmissions from query engine
- Open