Open Jellby opened 5 years ago
If gitlab sends a 'Retry-After' header in the 429 response, marge could then try to honour that
I don't know if it does (I can't reproduce it), but could marge assume a default of say 60 seconds?
The current throttling behaviour is documented in one of the comments of this thread.
I got this log:
as you can see, after waiting a couple of hours for the CI to finish (this is not a problem, the pipeline was running fine, it just takes a long time), marge got
429, 'Too Many Requests'
from gitlab.com and crashed.A better protection against this kind of error would be welcome, or maybe it's just a matter of increasing
waiting_time_in_secs
.