Test with multiple requests produces errors

I made some tests on 17.04.2023, with a set of 10 to 20 scenarios (I send attached an example), and I that is what I observed:

In the first run on Friday, with 11 new scenarios, the services shakemap_sampler, shakyground and deus executed sucessfully 11 times. The other services did not run, because the input for them did not change wrt the simulations done before this test.
Then I sent a request with the same 11 scenarios plus 10 new ones, and the services ran only for the new scenarios.
I repeated the run with the same scenarios, but changing the gmpes. Those who had a different gmpe executed shakyground and, if successfully, downstream services again. However, due to a typo I made in one of the gmpes, shakyground failed when executing with the gmpe with that typo (displaying "failed" in the job status), and the respective downstream services did not run either.
After fixing the typo and sending the requests again, I realized that none of the services were being called again.
Even after changing the scenarios to completely new ones, none of the services were executed.

riesgos / async