Skip to content

Comment on Heroku clone for Django?parent

Comments

How does it work? How do you handle persistence and parallelism?

The front end dispatches requests to a Varnish, which is in front of a router that queues up requests and sends them to available copies of the desired application. Persistence is Postgresql, and delayed and scheduled jobs are handled by celery over RabbitMQ.

My number one goal is to provide a deployment environment with as few additional restrictions on the developer as possible.

That sounds quite good! So most of the parallelism is on the view processing layer? Is the postgres setup parallelised as well?

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.