Do not replay search request when HLA validation failed

[benbelow]

since #707 we should be able to identify HLA lookup errors.

Currently these retry up to the retry limit (10 by default), causing a failure notification each time, and eventually ending up on the dead letter queue.

Suggestion: Instead, consider such messages "successfully processed". This would mean we only get one failure notification, and the messages will not end up on the dead letter queue, meaning we can reliably say that anything on the DL queue is a problem.

[benbelow]

Confirmed with @mmelchers and @WMDAJesse that this would be useful to WMDA

[benbelow]

TESTING

• Submitted a search request with invalid HLA
  • Confirmed that the message was removed from the search requests queue with no dead letter created
  • Confirmed that only one failure notification was sent
• Broke the SQL connection string for the matching database, to cause an "unexpected" error, and initiated a search with valid HLA
  • Confirmed that search request message dead-lettered, and so broken searches can be tracked via the dead letter queue