Queue Inspector MCP server

This is a small helper program that lets your AI assistant look inside job queues that live in Redis. A job queue is a waiting line of tasks your app needs to run, like sending emails or processing orders. If you work with a team that runs these queues, this makes it easy to ask plain questions about what is stuck, failing, or finished.

What this MCP server does

You ask your AI something like how many jobs are stuck or why one failed. The AI passes that question to this helper. The helper talks to your Redis queue and understands the way Asynq, BullMQ, and Sidekiq store their jobs. It reads the counts, the job details, and the error messages, then hands the answer back to your AI. Your AI shows you the result in the chat, in plain language.

Click to zoom

What you can do with it

• List every queue it can find, with the backend each one uses
• Count jobs per state, like pending, active, retry, or archived
• Page through jobs in one state and see their type and last error
• Open one job to see its full payload, attempts, and timestamps
• Retry a failed or dead job so it runs again
• Delete a single job permanently
• Run in read-only mode so nothing can be changed

Try asking your AI

• “How many jobs are stuck in retry in the default queue?”
• “Show me the last three failed jobs and why they failed”
• “What is in the archived queue right now?”
• “Retry the job with id 814cc556-04a8-4923-8de5-71a661c6063c”

What it gives back to you

You get back clear answers in the chat: counts per state, lists of jobs with their ids and types, and full details for a single job including its payload and error. When you retry or delete a job, it tells you whether the change worked. Everything comes back as structured data your AI can read and explain, not raw text you have to decode.

Before you start

Who it's for

Developers, support engineers, and operations people who already run Asynq, BullMQ, or Sidekiq queues and want to ask questions about them in plain words.