Salesforce Status Report – What Are We Up To With Spinnaker?
summary
Salesforce runs Spinnaker at ~400,000 executions per day and focuses on concrete fixes: implementing on-demand caching of accounts to avoid idle-account work, compressing pipeline execution json to fit SQL storage, and improving front50 SQL queries plus targeted Echo/Orca filtering. These changes cut pipeline-related traffic from 93.5% to ~13%, reduce front50 CPU from ~70 cores to ~5–10, and yield roughly a 60% cost reduction in compute for the rollout so far.
AI-written summary · about
chapters (11)
- 0:00Salesforce and Spinnaker intro
- 1:49Spinnaker microservices overview
- 3:32Early Spinnaker optimizations
- 4:55On-demand AWS caching
- 9:12Teams managing Spinnaker
- 10:58Pipeline config hotspot
- 15:08Query optimization results
- 17:13Disabling K8s caching
- 17:55Reducing Spinnaker costs
- 18:36Future Spinnaker improvements
- 22:24Q&A
related talks
27mSpinnaker at Salesforce: Enterprise-Level Multi-Substrate Continuous Delivery and Release Management
29mScaling Spinnaker for multi-region multi-cloud SaaS deployments
24mGrowing Pains: Taking off the training wheels at Salesforce
25mEnabling Diverse Application Use Cases using Spinnaker
45mEvolution of Spinnaker Internals & Operations
36mSpinnaker - Land of a 1000 Builds
33mPanel Discussion: Spinnaker Success - CD as a Service using Spinnaker
01h 04mOpening Keynote [Full]