Living without pre-production environments
summary
Claims production is the only environment needed and backs it with concrete tooling and lessons: uses slingshot for metric-driven CD, testcontainers to run identical CI and local integration tests, and dependable for org-wide library bumps; hits a hard ECS limit of 4000 containers and sees components handle ~2 million messages/sec; ships a new data‑ingestion platform by double‑dispatching into prod and validates via shadowing traffic to Kubernetes; a Kubernetes migration briefly breaks a cluster then succeeds after moving to a cells‑based architecture.
AI-written summary · about
chapters (6)
- 0:00Intro
- 0:20Introduction to Nicky
- 0:50Why do we not replicate to lower environments
- 3:28SkyScanner
- 6:20Cookie cutter
- 13:40Things to consider
related talks
45mLearning How to Pronounce Kubernetes to Production in 3 Months!
18mProduction Population Control: My Cattle are Rabbits!
22mYou don’t have a production environment
16mStaging Environments: A Blueprint for Pre-production
25mContinuous Delivery and Testing in Production
34mKubernetes from Dev to Prod at GoEuro
32mAutomating and Testing Production Ready Kubernetes Clusters in the Public Cloud
42m18 Months Before the Mast