On 2020-04-23 14:53:45 +0200 (+0200), Thierry Carrez wrote: [...]
reply: retcode (451); Msg: b'Temporary local problem - please try later' [...]
We've been witnessing occasional flurries of OOM events on the lists.openstack.org server, and this temporary delivery failure coincided with when one was happening so is almost certainly related. Whether it's a cause or an effect is unknown as we've been so far unable to track down what's actually triggering these sudden spikes in memory utilization, though given the volume of messages the server is handling on a constant basis I have serious doubts release activity is the problem. I suspect these release announcements just had unlucky timing. Normally this wouldn't cause a permanent delivery failure, but the release announcements job doesn't retry delivery on temporary failures. This is something which might be worth looking into adding if a better delivery guarantee is desired. -- Jeremy Stanley