[core-amqp] [Event Hubs] Long-term solution for OperationTimeoutError handling

[ramya0820]

Based on work being done as part of #4171, we have decided to tryout making changes in core-amqp to pass in OperationTimeoutInMs along with changes in for at least EventHubs before concluding fix for #4171 (so we can verify handling of all affected usecases).

As discussed offline, this issue is used to track changes required in core-amqp and EventHubs for the more right, long-term approach proposed for #4171

Once this concludes, depending on complexity involved, we can move ahead to update Service Bus (Track 2) if applicable.

[ramya0820]

Based on investigation done for #2891 (If OperationTimeoutError is to be retryable or not), some conclusions have been captured at #2891 (comment) that are applicable to identifying what are the affected operations that need to be updated.

Current efforts to test changes in EventHubs would help with it as well.

[ramya-rao-a]

It is now known that using this approach and ensuring that Event Hubs and Service Bus can continue to work as expected is not trivial.

Based on current bandwidth, we don't have time to explore this approach at the moment.

There are talks about having a pipeline for amqp libraries just like we do for http, in which case retry policy and implementations will be re-visited and re-factored at which point it would be expected for the retry() to implement its own timer.

Therefore, closing this issue.