Purging old EventStreams - when is it safe? What's the impact?

Hey everyone,

Is there a recommendation on the ideal length of time to preserve historical event streams and how large a window size to make on their purging?

The out of box default is (I think) 6 months maximum age and 1000 event streams per purge, but these numbers feel like they’re not well matched, particularly for a large region of a couple dozen appliances of more that can generate a lot of events per day.

Secondly, at what point is it safe to purge records from this table, and what is the impact on MIQ operation if these records are purged? Is 6mo set as the default for a particular reason, or is there any harm in shortening that right down (e.g. 7 days, 30 days)?

I ask because I have an event_streams table of some 7 million rows (!!) that needs pruning. I’d love to just TRUNCATE TABLE, but I figure there are some issues with truncating the table in a live region (e.g. lost events while they’re still in-flight) and so I figure I need something more refined.

If I purge all event streams older than, say, 7 days, is there any impact on MIQ’s operations?


1 Like