Datawarehouse queries could be long and run for a few minutes. So the goal is to seamlessly support long-running queries without the need of re-run those with UI actions.
As between cube.js and Metabase the connection is over http the connection could be terminated for many reasons. Also cube.js REST API can respond with continue wait so somehow we should tackle this and still wait for the results.
These are the default query timeouts (20 min for prod). I guess it's enough for us.
So the only thing we have to do is handle the Cube.js "continue wait" response.
Datawarehouse queries could be long and run for a few minutes. So the goal is to seamlessly support long-running queries without the need of re-run those with UI actions. As between cube.js and Metabase the connection is over http the connection could be terminated for many reasons. Also cube.js REST API can respond with continue wait so somehow we should tackle this and still wait for the results.