Questions taking a long time or failing

Get Help
1

Hello there,

I have noticed that some questions that use big tables (over 1 million records) and have over 10 columns aggregated tend to fail and it is not possible to run them again after they have failed. Message "something went wrong" (Picture below)
image

In the logs I am getting: "2023-08-22T16:45:40+01:00 INFO metabase.server.middleware.exceptions Request canceled before finishing."

The same query takes 10 seconds in the RedShift DB and over 40 seconds in Metabase.

So, the two issues I am finding are:

  1. Questions taking a long time in Metabase
  2. Some of these questions break and do not run again.

Is there solution for any of these issues?

Thank you!

2

I am using 0.47.

Everything looks good in AWS, no server issues, no ebs issue and no db issues.

If I start removing columns and joins till I am just doing
SELECT date, COUNT(*)
FROM table
GROUP BY date

It still retrieves the error of "Something's gone wrong"

3

There’s a regression in 47 that when the FE fails you can’t recover it unless you start the question again. Is fixed in 47.1, which version are you using?

4

Hi Luiggi :slight_smile:

I am using Metabase 0.47.0. I think it is the latest one.