Hi everyone, As OpenStack clusters grow larger, they hit scaling limitations in a number of components. To work around this problem, operators create multiple clusters. In recent visits to OpenStack users, I noticed a common theme around people interested in working to scale single clusters past their current limits. In the past we had various groups trying to identify bottlenecks or work on alternate technologies (RabbitMQ replacements): the Performance WG, or the "Large deployment" WG. However those groups had a bit of a wide focus, and/or are not active anymore. In Shanghai we'll discuss creating a "Large Scale SIG" that would tackle specifically those single-cluster-scaling issues. The idea is that by sharing performance analysis, it can help identify key bottlenecks. By pooling development resources, it can push fixes or new features to address those issues. I directly reached out to several people and organizations interested in working on that. Please join the session in Shanghai if you'll be present there and are interested in discussing this topic: https://www.openstack.org/summit/shanghai-2019/summit-schedule/events/24405/... -- Thierry Carrez (ttx)