Deployments We Can Believe In
summary
An internal Cloud Foundry pilot uses AWS, BOSH stemcells and BOSH manifest splitting to deliver a multi-tenant platform that runs for five months with 100 app instances and ~50 users managed by a 3-person team. Server utilization improves, user offboarding becomes trivial, and a claimed ATO cut from ~1 year to 2–3 days enables real production launches. Teams add CF CLI plugins, improve monitoring, and expose platform services like provisioned databases.
AI-written summary · about
chapters (9)
- 0:00Introduction to 18F and DevOps
- 1:28Legacy deployment challenges
- 3:27Attempting to improve workflows
- 5:36Selecting Cloud Foundry
- 7:50Leveraging BOSH and scaling
- 9:23The Cloud Foundry pilot project
- 10:27Results and developer feedback
- 12:40Key lessons and best practices
- 15:09Future goals and expansion
related talks
39mManaged CF-as-a-Service: An Operator's Cookbook
31mWinning in Production
33mGive Cloud Foundry to your Company
49mAutomation in Deployment on Hybrid Hosting and Private Cloud Environments
28mHow to Deploy CF and Other Bosh Releases on Bare-Metal Environment
29mThe Challenges of Taking Open Source Cloud Foundry to Production
25mDeploying Cloud Foundry with BBL, Bosh 2.0 and CF Deployment
43mHow Platforms Work