Comment 10 for bug 1816727

Jeremy, correct that this is not a bug in websockify. Apologies for not being clearer in comment 6. I should have said, "I did some local testing and discovered that this does not appear to be a bug or issue in websockify and that the fix can be handled entirely in nova."

On your second question, I'm not 100% sure. The creation of the ComputeAPI object happens during the handshake flow, so I would think even with the attached patch, one could cause similar resource consumption by just opening a bunch of connections to the socket and not closing them.