Failed Messages can be duplicated when attempting to retry

[jpalac]

Describe the bug

Symptoms

Retried messages can be processed multiple times at the originating endpoint.

Who's affected

Anyone using ServicePulse to retry messages who selects to retry the message more than once. For example, The message may be submitted for retry as a batch of related messages, and whilst that's processing, it's also resubmitted for retry as an individual message.

Root cause

BatchDocumentId being updated in the RavenDB on a subsequent retry attempt for the same messageId.

## Description

If a failed message is retried in ServicePulse and the user refreshes the page before the retry is acknowledged, or submits the message as a batch retry and then straight away as a single retry, then ServicePulse allows for the message to be retried multiple times.

Expected Behaviour

The system prevents the message from being retried if it's currently queued to be retried

Observed Behaviour

The user can retry the message in ServicePulse as part of a batch, and then whilst that's processing, immediately submit the same message for retry as an individual retry.

Backported to:

• Fix batch message deduplication logic to not update batch on already batched message - 5.11 #4795
• Fix batch message deduplication logic to not update batch on already batched message - 6.3 #4796

[PhilBastian]

fixed by #4789

[jpalac]

Versions 6.3.2 and 5.11.5 have been released.