How does GCLB connection drainning apply to Cloud Run when Cloud Run + Serverless NEG is used as a backend? Also, when using http or http/2 as a connection between LB and Cloud Run, is there any difference in behavior between the two?
I am using Cloud Run in the above configuration, and I am getting 503 errors when the container drops, so I thought it might have to do with Connection Drainning in GCLB.
