CHOOSEDevtools & code
DLQ Inspector with Jev Integration
HalxDocs@HalxDocs
DLQ Inspector analyzes and classifies failed messages from dead-letter queues, now optionally using Jev by TypeSafe AI for classification when existing rule-based systems are insufficient.
SignalThe flow is: headers → policies → Jev → rules So if the deterministic checks can't confidently classify a message, Jev gets a shot at it and returns a structured decision with calibrated confidence.
I just integrated Jev by TypeSafe AI into DLQ Inspector.
First, what is DLQ Inspector?
When a message fails processing in a system like RabbitMQ or Redis Streams, it can end up in a dead-letter queue (DLQ).
The problem is: you can't just replay everything.
Some messages are safe to replay.
Some need to be fixed first.
Some should never be replayed.
And some don't have enough information to make a confident decision.
DLQ Inspector helps engineers inspect, analyze, classify, and safely recover those failed messages.
Now it can use Jev when the existing rule-based classifier gets stuck.
The flow is:
headers → policies → Jev → rules
So if the deterministic checks can't confidently classify a message, Jev gets a shot at it and returns a structured decision with calibrated confidence.
A few things I cared about:
• BYOK
• Metadata-only input
• Jev is opt-in
• Silent fallback to the existing rules if Jev isn't available
• Existing safety controls stay in place
• Nothing gets replayed automatically
Enable it with:
dlq jev enable
DLQ Inspector is open source and supports RabbitMQ + Redis Streams.
If you're dealing with failed async jobs in production, give it a try.
github.com/HalxDocs/dlq_inspector