networking – Azure Service Bus TimedOut exception when running within K8S instance but not in docker on node host

I’m getting a “ServiceBusCommunicationException” reason “Connection timed out ErrorCode: TimedOut” with an Inner Exception of “SocketException” reason “Connection timed out” when running a .net core 3.1 application in a docker container running in K8S.

K8S is running on Ubuntu 20.04 LTS

When I run the container directly on the host of the node itself (i.e. not in K8S) the connection is fine and it processes messages ok but in K8S I get the above error. There is no firewall running on the host. The containers are based on the mcr.microsoft.com/dotnet/runtime:3.1-buster-slim image. There is nothing else added to this or any other tweaks made.

This suggests a K8S issue, possibly networking, with the AQMP port itself.

K8S reports no network policies in place.

As part of the same container I have a chron job that runs that makes an http GET to the google URL. This continually reports OK (200).

Therefore (at least for HTTP) the pods are communicating with the internet fine.

If I change the authentication credentials to something “bad” I get an unauthorised error … so it does appear to be routing ok for authorisation but not when actually connecting to do some work.

I’m not sure where to go with this. Any advice?