Dead letter queue
A holding queue for messages or jobs that could not be processed successfully after their retries were exhausted, preserving them for later inspection.
In more detail
A dead letter queue is where messages or jobs go when they cannot be processed successfully even after retries. Instead of discarding a persistently failing item and losing it, the system moves it to a dead letter queue where it is preserved. This turns a silent loss into a visible, inspectable failure.
The pattern is essential for reliability at scale. It keeps a stuck item from blocking the rest of the queue, and it gives an operator a place to look for what failed, understand why, and decide whether to fix and replay it. A system without one tends to either jam on bad items or lose them without a trace.
Where this shows up at Ceven
The principle behind a dead letter queue, never lose a failed item silently, is exactly what Ceven's exception handling and audit trail provide at the workflow level. A step that cannot complete after retries is surfaced and recorded rather than dropped, so an operator can see what failed and why instead of discovering the gap later.