Lessons Learned from building out hyper scale cloud services using docker
summary
Oracle builds an internal hyper-scale configuration and leader-election service backed by Coherence, running on bare-metal OCI and orchestrated with uses Docker Mesos/Marathon and Kubernetes; the production system has nine separate services delivered by nine services delivered by five geographically separated teams and supports multi-tenant tenant load balancers, dynamic reconfiguration, and Jepsen chaos testing. CI/CD automates verification, parallel 10-minute acceptance shards, staged upgrades and human-gated stress review; container sizing and JVM memory inside Docker required manual tuning but the platform is operational.
AI-written summary · about
chapters (24)
- 0:00Intro
- 0:58Agenda
- 1:57Objectives
- 2:29The use case
- 3:29Service design goals
- 4:55Service design principles
- 7:44Technology stack
- 9:22Architecture
- 11:13Platform Components
- 12:22Orchestrator
- 15:03Runtime Architecture
- 15:47Testing Strategy
- 18:33CICD Pipeline
- 26:33Lessons Learned
- 28:28Microservices
- 30:28Docker in Java Apps
- 32:38Docker Startup Script
- 37:53Dynamic reconfiguration
- 39:18How it works
- 40:53Versioning
- 43:08Formation
- 45:11Agile Testing Pyramid
- 47:29Diagnostics
- 50:19Summary
related talks
29mDocker at Scale
39mExperiences and Lessons Learned Building a Multi-Cloud Spring Boot Backend
39mHow we learned to love the Data Center Operating System
40mGoes Beyond CI to Production Scale PaaS with Docker
34mKeynote: Dockerizing legacy apps
01h 18mEmbracing Docker simplicity - while harnessing platform power
17mThe Twelve Days of DevOps
36mDocker Inside/Out