Database queries sometimes take a lot longer to complete than the average developer would expect. Seemingly innocent actions such as searching can pro

Django + Postgres: The Hunt for Long Running Queries

submited by

Style Pass

2024-10-10 18:00:05

Database queries sometimes take a lot longer to complete than the average developer would expect. Seemingly innocent actions such as searching can produce very complex SELECT statements which take what feels like an eternity to finish. If you get a few users performing these heavy queries in parallel, your app’s performance may begin to suffer.

In this article we reflect on how we at Intempus implemented some basic tooling to help prevent getting into the scenario described above. In particular, we describe a way to detect long running active queries and also a way in which we can set a timeout on a query-by-query basis. We speak in terms of implementing these ideas in Django and Postgres, but most concepts should be transferable to other frameworks and databases.

I work as a Software Engineer at Intempus , a time-registration software company based in Denmark. I help develop the backend of the app and website, which is built from Django on top of a Postgres database.