Redshift cancel

This issue https://github.com/metabase/metabase/issues/2399 is apparently fixed by this merged PR 6848 https://github.com/metabase/metabase/pull/6848 suggests that “cancel” has been fixed for “databases that support it” but it doesn’t work on Redshift.

There are also these related issues:


And this PR

I know that it doesn’t work because I can open the AWS Console and see the query still running.

Other database query tools, for example DBeaver which uses the JDBC driver, are able to cancel queries.

I have the ongoing issue of users cancelling and re-running the same query multiple times and filling the concurrency slots, and continually explaining to users that the cancel button does not work. It’s not much good knowing the button doesn’t work.

Are there any plans to fix it?

On a related note, the autorunning of queries without parameters is not great when what you want to do is open a question and modify it slightly to create a new question.

We’ve been using this hack

  • add an enable_autorun parameter of type bool with no default value
  • Add this in the where clause:

WHERE True = {{ enable_autorun }}

This is pretty gross, it would be nice to configure default to not-autorun and have a go button, either as a default config for all reports or for a specific report.