Is there some sort of daily maintenance that leads to closed connections?

We seem to experience a ConnectionClosedException once every 24 hours, every single day. Is this part of some sort of regularly-scheduled maintenance?

Is the proper way to handle this to reconnect, and resubscribe? If so, how do we ensure there’s no gap between the end of live processing before the connection closes, and new events afterward? We are using “$all” to process all events to push into a NoSQL store. Does tracking the last commit position make sense, and simply starting from there, or am I missing something there?

We don't do anything along these lines though the environment might have things.

What you was is aubscribetoallfrom that will in general prevent you
from losing any events (and the catchupsubscription handles the logic
for you)

Noted, thanks.

BTW, trying to resubscribe without reconnecting (didn’t initially track/notice the different drop reason) was what surfaced that mysterious FormatException I’d reported in a support ticket. I would expect an application specific exception like ConnectionNotOpenException or something else descriptive … that FormatException had me, and client IT, hunting a lot of red herring.