Mikhail Panchenko - Spinnaker Summit 2019
summary
Platform handles 1.3 billion events per minute (about 22 million/sec), supports ~500 engineers and more services than people, and still runs a Marathon fleet for stateless apps while rolling out a production multi-cluster Kubernetes tier for stateful, complex services. Built a declarative registry called Grand Central so repos describe deployment intent, uses CRDs/operators to manage stateful clusters like NRDB, and integrates Spinnaker + Kayenta canaries for visibility — the hybrid approach is live and incrementally working.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 1:06Flickr
- 2:55Relic
- 3:50Marathon
- 5:07Grand Central
- 6:20Software
- 6:37Platform
- 7:05Migration
- 8:08Deployment
- 8:43Community
- 9:20Onboarding
related talks
33mPanel Discussion: Spinnaker Success - CD as a Service using Spinnaker
26mSpinnaker Summit
36mSpinnaker - Land of a 1000 Builds
30mSpinception: Deploying Spinnaker with Spinnaker
25mEnabling Diverse Application Use Cases using Spinnaker
18mOvercoming the challenges of using Spinnaker as a multi-tenant CD service
19mKeynote: Spinnaker - The Future of Your Enterprise CD? Three Reasons Why
34mSpinnaker Technical Oversight Committee Panel