Keynote: 5 Years of Cloud Foundry at Rakuten
summary
A five year production Cloud Foundry deployment runs on 5,000 virtual servers managed by a team of seven engineers and evolves from a forked platform toward upstream CF v2 migration. Architecture uses BOSH, Concourse, VMware + OpenStack directors to run one logical CF across two clouds, enforces env isolation with router groups and ACLs, and implements placement pools. Logging and metrics funnel into a Kafka-based log pipeline with three day retention and blobstore backup; DNS flakiness and lost logs drive monitoring and tooling changes.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 0:30Harpus
- 1:51First Import
- 4:43Monitoring System
- 5:56Locked Pipeline
- 6:49Problems
- 7:41Deployment
- 9:22Deployment Environments
- 12:04Kafka Pipeline
- 16:57Feedback
related talks
23mFive Years of Cloud Foundry at Rakuten
21mKeynote: Cloud Foundry at Rabobank after 10 months of real use
11mKeynote: Steampunk Meets Cyberpunk: The Makings of a Cloud Native Application Platform
33mCloud Foundry Weekly: Live at CF Day Germany
17mDeployments We Can Believe In
29mThe Challenges of Taking Open Source Cloud Foundry to Production
15mThe Cloud Foundry Renaissance
48mCloud Foundry - your PaaS on OpenStack