Memory leak with /migrate

[celian-garcia]

What did you do?

I made a script to migrate all dashboards from Amadeus Grafana instance to Perses instance. It works well but at some point, when it arrives to some dashboard there is a big memory leak.

I'm doing the migrate for around 1k dashboards. And it seems to leak when it arrives to some dashboards with row repeat. I'm not sure yet it's linked but I will dig further to isolate the dashboards and see if I can reproduce with a few examples.

Of course I cannot share the dashboards

What did you expect to see?

Memory correctly released by the Garbage Collector

What did you see instead? Under which circumstances?

System information

Linux 6.5.0-28-generic x86_64

Perses version

e9dddceb0fdf640e9b0ebdb278b0dde8b5f49216

Perses configuration file

No response

Logs

No response