There is a bug in crust, not bad, we can see it and have been tracking this afternoon. Instead of logging a particular error (too many timers) it disconnects the node instead. We are able to sort the symptom easily enough, but are trying to find out why too many timers are established before committing to a fix. At the moment this is causing connection loss, but we are leaving it right now to analyse the network under this loss as it would be a weird attack to be able to do this, but it’s good info.
So not a big worry, however we do need to patch it.