Describe the bug

Description

Using service pulse, and "Edit and Retry" on a failed message, doesn't trigger the event MessageFailureResolvedByRetry. If doing a "regular" (no edit) retry, the event is triggered.

The message that was retried with an edit doesn't contain the following headers:

And according to the documentation:

A MessageFailureResolvedByRetry event is published when an audit message is detected that contains ServiceControl retry headers. The primary instance subscribes to this event on order to be able to mark the failed message record as successfully retried.

I know that it's not advisable to edit the message body as it's altering the "history", but sometimes some invalid characters not picked up by our validation yet (not known to BE a validation error yet as it's a third party system that holds all the rules) is in some part of a message and for us it's easier to edit the message and retry while updating our validation to check for the newly found "bad" character.

We have subscriptions on MessageFailed to show our application administrators that something has failed and is currently (probably) being handled by a developer. When the retry is done we also subscribe to MessageFailureResolvedByRetry to automatically show the admins that the issue has been resolved. This doesn't work automatically if the developer had to edit the message prior to the retry.

Expected behavior

Raise the event MessageFailureResolvedByRetry no matter how the failed message is retried.

Actual behavior

MessageFailureResolvedByRetry is not raised if the failed message was edited prior to retry.

Versions

Please list the version of the relevant packages or applications in which the bug exists.

ServiceControl: v6.7.3 (and earlier)

Steps to reproduce

Edit the message body of a failed message and retry.

Relevant log output

Additional Information

Workarounds

Possible solutions

Additional information