Prefetched messages cannot be drained when pausing receiving

[SeanFeldman]

Description

Prefetching improves the overall throughput and is a desired setting for most high-load scenarios.
Unfortunately, a design flaw prevents prefetched messages from being processed when a receiver is passed. An example would be using ServiceBusMessageProssesor, which can be paused and restarted. When a message is prefetched, and the processor is paused, the message's lock starts ticking, but it cannot be received by the processor. This causes the message's lock to expire and the message to be retried later. The assumption that a retry would fix the issue does not apply to all scenarios, especially when a processor could be stopped and started multiple times. This situation causes many messages to end up in the dead-letter queue due to repeated LockLostException due to the message being pre-fetched. And the more aggressive the prefetch, the worse this problem becomes.

Actual Behavior

1. Processor paused
2. No prefetched messages are drained
3. Processor restarted
4. LostLockExceptions thrown with some messages eventually ending in the dead-letter queue due to exceeded maximum number of deliveries

Expected Behavior

1. Processor paused
2. Prefetched messages are drained
3. Processor restarted
4. No LostLockException

[EldertGrootenboer]

Thank you for suggesting this feature. We have opened an investigation task for this in our backlog, and will update this issue when we have more information. To help us get this prioritized, it would be helpful to see others vote and support this feature, as well as explain their scenarios.