How to Use Durabull
Connect your Redis instance, find your queues, and investigate your first failed job.
Durabull connects to Redis to inspect and manage the jobs your BullMQ application creates. It does not run your workers. You can explore queues without changing worker code; throughput charts need BullMQ metrics enabled on workers.
Get connected
- Open Durabull Cloud, install a desktop app, or follow the self-hosted installation.
- Sign in and create or select an organization when authentication is enabled. Authless mode creates a local organization automatically.
- Select a Redis connection, or add one in Settings → Connections when UI management is enabled.
Match the Redis database and BullMQ prefix used by your application; the default prefix is
bull. - Open Queues. An empty Redis instance has no queues until your application creates them. If you expected queues, use Connection Troubleshooting.
The Redis URL must be reachable from the Durabull API. For desktop, that is your workstation; for Docker, it is the container; for Cloud, it is the hosted service.
Investigate a failed job
- Confirm the selected organization, connection, and environment.
- Open a queue with failed jobs, then filter its jobs by Failed.
- Open a job and read the failure reason, stack traces, data, and logs. Workers must write
logs with
job.log; console output does not automatically appear here. - Fix the cause in your application or its dependencies. Before retrying, consider whether the previous attempt may already have performed part of its work.
- Retry the job and watch its state and new logs. If it stays waiting, confirm that a worker is listening to the same queue and prefix.
Use a development queue for your first drill. See Job Lifecycle and Debugging for editing, retrying, and removing jobs.
Find the right workspace
| Workspace | Use it to | Read next |
|---|---|---|
| Queues | Compare current job counts and spot backlogs | Queues Dashboard |
| Queue detail | Inspect jobs, metrics, and queue controls | Queue Operations |
| Analytics | Compare throughput and fleet signals | BullMQ Native Metrics |
| Scheduled Jobs | Create and manage recurring job templates | Scheduled Jobs |
| Workers | See Redis clients and connection activity | Workers Topology |
| Redis Keys | Inspect keys, values, and expiry | Redis Key Explorer |
| Alerts | Monitor failures and Redis pressure in the background | Redis Health and Alerts |
Queue pause stops workers from starting new jobs; active jobs can continue. Purge and delete have confirmation steps. Obliteration is a separate destructive action without those safeguards. Read Queue Operations before removing queue data.
Make future incidents easier
- Use consistent queue and scheduler names.
- Write structured job logs.
- Route alerts to webhooks or Linear.
- Use Troubleshooting when a queue is missing, jobs remain waiting, or integrations fail.