The user should be able to configure automatic fixes

[bzizou]

For example, in the JDL, we could have:

"action_on": {
"timeout": "ignore|resubmit|blacklist",
"walltime": "ignore|resubmit|blacklist"
},

with ignore=fix the event, resubmit=fix the event and resubmit, blacklist=disable the cluster until manual fix

[bzizou]

for RUNNER_SUBMIT_TIMEOUT, we can also have a "retrieve" option: try to retrieve the submitted jobs by searching jobs submitted in the time interval with the name of the campaign... or tagging jobs (how?)