Skip to content

Odoo Job Worker

High-performance, SQL-driven background job engine for Odoo.

Odoo Job Worker replaces traditional cron-based or memory-resident job queues with a PostgreSQL-native architecture built on FOR UPDATE SKIP LOCKED and LISTEN/NOTIFY. Jobs are durable rows in the database — no broker, no Redis, no external dependencies beyond PostgreSQL.

Features

  • SQL Pull Architecture — Workers acquire jobs atomically with SELECT ... FOR UPDATE SKIP LOCKED. No double-execution, no message loss.
  • Instant Wakeup — PostgreSQL LISTEN/NOTIFY triggers workers the moment a job is enqueued. No polling delay.
  • Per-Channel Throttling — Concurrency limits and rate limiting (jobs/second) per channel via queue.limit records.
  • Job Graphs — group() for fan-out parallelism, chain() for sequential pipelines, on_done() for callbacks.
  • Heartbeat & Timeout — Workers send heartbeats during execution. Stale jobs are automatically recovered. Per-job timeouts are enforced.
  • Familiar API — with_delay(), delayable(), group(), and chain() — compatible with OCA queue_job patterns.
  • Multi-Database Runner — A single runner process discovers databases and manages per-database worker threads.
  • Dashboard & Alerts — The job_worker_monitor module provides a real-time dashboard, metrics collection, and configurable alert rules.

Modules

Module Description
job_worker Core queue engine — job model, worker, delay API, channel throttling
job_worker_monitor Dashboard, metrics, and alert rules for queue operations
job_worker_demo Interactive demo companion for exploring the queue system

Requirements

  • Odoo 19.0
  • PostgreSQL 12+ (including PostgreSQL 18)

Quick Start

# Enqueue a job with one line
partner.with_delay(channel="exports", priority=5).write({"name": "Updated"})

See Getting Started for installation and your first job.

Architecture Overview

graph LR
    subgraph Odoo Application
        A[with_delay / delayable] -->|INSERT| B[(queue.job table)]
    end
    B -->|NOTIFY| C[QueueJobRunner]
    C -->|Spawns per-DB| D[QueueWorker]
    D -->|SELECT FOR UPDATE<br/>SKIP LOCKED| B
    D -->|Executes| E[Job Method]
    E -->|UPDATE state| B

Worker / Runner Relationship

graph TD
    R[QueueJobRunner] -->|Discovers databases| R
    R -->|Spawns thread| W1[QueueWorker DB1]
    R -->|Spawns thread| W2[QueueWorker DB2]
    R -->|Health check| W1
    R -->|Health check| W2
    W1 -->|Acquires jobs| P1[(PostgreSQL DB1)]
    W2 -->|Acquires jobs| P2[(PostgreSQL DB2)]