QRDS slow and timing out

Issue Summary

We are using self-hosted redash. I am using QRDS to do the following:

SELECT
q1.id
FROM cached_query_1 q1
LEFT JOIN cached_query_2 q2 ON q1.id = q2.id
WHERE q2.id IS NOT NULL

Both queries take about 3 seconds to run on their own (outside of QRDS). Both contain a single column of ids and have about 40K results.

Cached QRDS runs for about 20 minutes and eventually times out.

I’ve tried using query_1 and query_2, rather than the cached results and had the same result.

Finally, I tried:

SELECT *
FROM query_3
LIMIT 10

outside of QRDS, query_3 takes about 1 second to run and has 18 results. It took about 10 minutes to run in QRDS but, unlike q1 and q2, it ran successfully.

There are no notes in q1, q2 or the QRDS query as I saw that was an issue for someone else.

Any suggestions? If this could be related to our configuration, any guidance you can provide on what to investigate would be helpful too.

Thanks!

Technical details:

  • Redash Version: 10.0.0 (9c928bd1)
  • Browser/OS: Chrome (Version 96.0.4664.110) / Mac 12.1
  • How did you install Redash: standard Redash Docker image, no modifications

Sorry to just get to this question. How much memory does your Redash instance have? It seems like you might have constrained resources.