I am not sure if anybody else has seen or been impacted by this but we have recently completed an analysis of am year old support case that yielded that the root server has a hard cap of number of registration-based connections it can open to database at the same time. While that in itself is not unexpected, the unexpected and concerning aspect of is that once it happens Root Server enters this “hung” state (my term, not Support’s) where NOTHING works:
Console - if you are already in, you can’t submit actions or do anything; if you are not in, you can’t login; etc
WebUI - you can login but you can’t manage anything related to clients
Web Reports - cache updates are impacted.
Platform RestAPI - no longer works
and this state lasts for several hours (Support has claimed - we let it go until 2.5 hours and it certainly didn’t recover in that time frame). A simple “restart” of the root server does get it cleared but the overall lack of stability and the length of the outages are extremely troubling to us, hence, we submitted an idea for the improvement required (as per Support/L3 this will require redesign of the Root Server registration plugin). Anyway, if you agree that this is something that needs to be addressed and may impact you, please vote. Thank you.