I was helping troubleshoot an issue with Magnum recently and found out that there was a minor issue with how they handle RPC connections, and decided to do an audit of all OpenStack projects and found a bunch that are likely to be experiencing similar issues. The gist of the issue is that most of these projects creates a new rpc transport on every API call. If someone has some time over to help review these fixes, I would be grateful. https://review.opendev.org/c/openstack/magnum/+/770707 - Re-use transport for rpc calls https://review.opendev.org/c/openstack/magnum/+/770720 - Re-use transport for rpc server https://review.opendev.org/c/openstack/blazar/+/771110 - Re-use rpc transport https://review.opendev.org/c/openstack/solum/+/771111 - Re-use rpc transport https://review.opendev.org/c/openstack/murano/+/771341 - Use common rpc pattern for all services https://review.opendev.org/c/openstack/watcher/+/771381 - Use common rpc pattern for all services I tried to use a standardized approach with all of these, but feel free to add any feedback. -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.openstack.org/pipermail/openstack-discuss/attachments/20210123/aced55e5/attachment.html>