What happens when an Oban queue quietly stops doing its job? This guide groups production problems into three useful categories: systems that are totally stuck, queues whose backlog keeps growing, and jobs left hanging in execution. It shows how to detect each before customers report trouble, using Obanβs telemetry, logs, periodic sentinel jobs, and metrics from Oban.Met.
The practical advice includes tracking available and completed jobs by queue, exporting metrics to a StatsD-style collector, and setting worker timeouts and Pro job deadlines. It also covers common causes of stalled processing, including broken leadership and notifications, and explains why a healthy-looking application can still be clogged by long-running jobs. The techniques apply to Oban for Python too, with a caveat: querying metrics through Met requires an Elixir node.
continue reading onoban.pro
β οΈ This post links to an external website. β οΈ
If this post was enjoyable or useful for you, please share it! If you have comments, questions, or feedback, you can email my personal email. To get new posts, subscribe use the RSS feed.