I ve noticed that a panicking Durable
I've noticed that a panicking Durable Object causes the Worker that's calling it to just hang indefinitely. I'm guessing that the DO just aborts and somehow the socket connection isn't even getting cleanly shut down or something, so the worker doesn't even know that anything is wrong until it times out. Is there a way to change this behavior?
1 Reply
I guess it's worth pointing out that I've only observed this behavior with local dev; maybe it's different on prod, but I'm still evaluating DOs before getting a paid account
I'm also concerned about what this means for a DO that has a lot of websocket connections.. I guess a panic means the whole thing dies and everybody's got to wait for a timeout to realize that they lost their connection and then reconnect