3. Check Queue Health

Purpose: Check queue depths, dead letter queue counts, and processing rates to confirm the pipeline is healthy.

Prerequisites:

  • Bun installed with bun install completed
  • Upstash Redis credentials in .env.local
  • Admin app running (optional, for dashboard view)

Cost / Duration: $0 | ~2 minutes

Prompt

Run a pipeline health check to inspect all queue depths, DLQ counts,
and processing rates.

Run the diagnostic script:
```bash
bun scripts/diagnose-pipeline.ts
```

This reports queue depths, DLQ counts, processing rates, and any stalled entries.

Optionally, check the admin dashboard for a visual overview:
- Production: https://admin.eko.day/queue
- Local: http://localhost:3001/queue

**Healthy state looks like:**
- All queue depths < 100 (steady state)
- DLQ counts at 0 or near 0
- No stalled entries reported

**Action items if unhealthy:**
- High queue depth not decreasing --> workers may not be running (see Start Workers)
- DLQ count growing --> messages failing repeatedly (see Drain Dead Letter Queues)
- Stalled entries --> worker processing hung (see Recover Stuck Queues)

Verification

  • Diagnostic script runs without connection errors
  • Queue depths are within normal range (< 100 for steady state)
  • DLQ counts are 0 or near 0
  • No stalled entries reported

Back to index