Hey @<1798887585121046528:profile|WobblyFrog79> , yes testing this locally it does seems to solve the issue, thank you.
I will test it in our env.

On a different issue, have you any solution on how to make the agent listen to multiply queues?
On the helm it is written :

  # -- ClearML queue this agent will consume. Multiple queues can be specified with the following format: queue1,queue2,queue3

But this does not work as the agent will read them all as one queue