Resolved
--
We had some Ceph MDS nodes reconciling on the wrong version, which unfortunately led to a massive amount of connection failures and a cluster-wide blacklist of internal IPs. it should be all sorted now. Apologies for the delay while we navigated these, everything should be stable now. If your servers are still not working, please perform a hard refresh (Ctrl+F5), and if you're still experiencing problems, please reach out to support and we'll be happy to look into it and get this resolved. Thank you for your patience!
