Forklifting Django: Migrating A Complex Django App To Kubernetes
summary
Claims a full “forklift” of a Django monolith into Kubernetes is practical and repeatable. Uses multi-stage Docker builds, keeps Postgres on RDS and RabbitMQ on Cloud AMQP, then moves from Helm to a custom Kubernetes operator (built with controller-runtime in Go) that enforces ordering with a small state machine so migrations run once. Adds gitops via Argo CD, a CLI for common ops, and notes Python2 limits prevent using celery BX so celerybeat stays stateful. The system works in production.
AI-written summary · about
chapters (10)
- 0:00Start
- 0:23DEPLOYMENT IS HARD
- 2:22MICROSERVICES?
- 3:30ROUND ONE
- 7:05HELM AND OTHER HAT-BASED TECHNIQUES
- 8:45SECTION THREE
- 8:59CRDS, WATCHES, CONTROLLERS OPERATORS
- 10:22RECONCILE REPEAT
- 20:42DRIFT
- 21:00GITOPS!
related talks
30mForklifting Django: Migrating A Complex Django App To Kubernetes
38mEnd-to-End Django on Kubernetes
49mFundamentals of Kubernetes for Django developers
44mMigrating a Monolith to Kubernetes
32mDeploy Django: GitOps & Kubernetes Made Easy
44mDeploy Django: GitOps & Kubernetes Made Easy with Calvin Hendryx-Parker
26mDeploy Python Apps on Kubernetes and Prepare for Scale
31mSimplifying Python Web App Operations: Automating K8s Ops with Open Source