Retries are insufficient

[systemmonkey42]

Is your feature request related to a problem? Please describe. When connected to a service in my kubernetes cluster, if a pod is upgraded, it will be shut down and be replaced with a new version.

When connecting to a websocket service which is shutdown the ingress provider returns a 503 immediately.

If this takes more than a few seconds, the websocket retries are exhausted and I need to manually reconnect.

Describe the solution you'd like Configurable or infinite retries

Describe alternatives you've considered All alternatives are manual.