I observed it via the 'sm' command using telnet. It is back in the same state now - see below.
The most significant changes I've made recently have been to add logging via the internal webserver (which is an amazing feature, btw) - could that be related?