Variable going crazy after grafana migration

[AntoineThebaud]

Context

I encountered a case with a specific dashboard where doing grafana migration will cause to have some variables recomputing endlessly.

See video:

2023-10-27_14h38_27.mp4

These 2 variables have in common that they are relying on a query that uses $__range as time range argument to a promQL function. If I replace these by hardcoded timeranges, the issue stops. See video:

2023-10-27_14h42_00.mp4

I didnt time to check if the same is happening with a dashboard created from scratch, but I don't expect this to be linked with the fact this is a migrated dashboard. However maybe the issue is happening only when you have a quite big amount of data queried behind the scenes? (Which is the case in the example provided).

Perses version

main-2023-10-26-d44ceabd

[Gladorme]

I am checking if I can fix the issue. For now, I see that if you close all group panels of the dashboard, it's stopping the issue 🤔