Monitoring

The job_worker_monitor module extends the core job_worker with a dashboard, metrics collection, and configurable alert rules.

Installation

odoo -d <db_name> \
  --addons-path=/path/to/odoo/addons,/path/to/job-worker-modules \
  -i job_worker_monitor \
  --stop-after-init

job_worker_monitor depends on job_worker and mail (for alert notifications).

Dashboard

Navigate to Queue Jobs > Dashboard for a real-time overview of your queue:

• System health — queue depth, active workers, failed jobs
• Performance — failure rate (1h), throughput (1h), P95 duration (1h)
• Needs attention — stuck jobs, aged pending, retry exhausted, top failure type
• Drill-down actions — each KPI opens a filtered queue.job list

Metrics

The module collects metrics via a scheduled cron job (queue.job.metric). Metrics include:

Metric	Description
Jobs Completed	Count of done jobs in the snapshot window
Jobs Failed	Count of failed jobs in the snapshot window
Average Duration	Average duration of completed jobs
P95 Duration	95th percentile duration of completed jobs
Queue Depth	Current pending + waiting jobs
Active Workers	Distinct workers with fresh heartbeats

Duration Buckets (Job View)

Jobs are categorized by execution time using queue.job.duration_bucket:

Bucket	Duration
instant	< 1 second
fast	1–10 seconds
moderate	10–60 seconds
slow	1–5 minutes
very_slow	> 5 minutes

Alert Rules

Configure alert rules at Queue Jobs > Alert Rules to get notified about queue problems:

• Failed job threshold — Alert when failed job count exceeds a limit
• Queue depth threshold — Alert when pending jobs exceed a limit
• Stale job detection — Alert when started jobs have old heartbeats
• P95 duration threshold — Alert when tail latency degrades
• Average wait time threshold — Alert when jobs wait too long before start

Alerts are delivered via the Odoo mail module (email notifications, Odoo inbox).

Operational Tips

Investigating Failed Jobs

1. Navigate to Queue Jobs > Jobs and filter by State = Failed
2. Open a failed job to see the Exception Info field with the full traceback
3. Use Requeue to retry the job (resets attempts and clears the error)
4. Use Open Related Record to jump to the target record

Bulk Operations

The job list view supports bulk actions:

• Requeue Selected — Reset selected jobs to pending
• Set to Done — Mark selected jobs as completed
• Set to Failed — Mark selected jobs as failed

Troubleshooting

If jobs are not running:

1. Confirm the worker process is running — Check systemd/Docker logs
2. Check scheduled_at — Jobs with a future scheduled_at wait until that time
3. Check channel limits — Verify queue.limit records allow sufficient concurrency
4. Check PostgreSQL connectivity — Worker needs a direct database connection
5. Inspect heartbeats — Stale heartbeats indicate worker crashes; jobs will be auto-recovered