Table - query_queryexecution is 7gig in size

#1

IS there anyway to archive data from query_queryexecution so that it’s size doesn’t keep ballooning? Or can this execution history be disabled?

#2

whoa.

how many rows are in it?

we’re going to start summarizing + rotating those. we use query timings all over the place and are going to be using the history for a host of other features. you can delete anything older than say 2 weeks without any major harm, though it might mean some of our “smarter” features start acting stupid in some situations.

#3

Hi,

We have 4897543 rows at present…and growing.

Seems to be excessive, but it’s a busy/popular tool.

Cheers Joe.

#4

that should still not add up to 7 gigs =/

what application db are you on?

#5

It’s on a Mysql, innodb engine, row_format=compact.

It was migrated from H2 to MySql.

The count is now 4906272

#6

FYI I opened an issue for this at https://github.com/metabase/metabase/issues/4155

Discussing causes + fixes there.

#7

And this is still an issue.
The ‘query’ table can grow out of proportion with no mainstream way to manage it.

#8

Hi @CharlesB
Yes, and issue 4155 is still open. You can truncate the table or run a nightly script to maintain more fine grained control.
Which version of Metabase? And are you using Postgres or MySQL for metadata?

#9

Metabase v0.31.2 (in kubernetes) with Postgres.