How Salesforce is using ArgoWorkflows for provisioning add-ons
summary
Salesforce migrates add-on provisioning from Spinnaker to Argo Workflows, converting pipeline templates into Kubernetes CRD workflows that run Helm and Terraform and store execution state in Aurora/RDS and artifacts in S3. The platform runs an Argo controller per environment, uses reusable workflow templates and native retries, and implements memoization and Prometheus metrics. Early results: 4 of 19 add-ons converted, ~20% provisioning time improvement for logging/metrics, and ~50% faster inner-loop velocity for dev testing.
AI-written summary · about
chapters (9)
- 0:00Salesforce and Hyperforce intro
- 4:00Kubernetes scaling challenges
- 8:17Spinnaker add-on provisioning
- 11:30Argo Workflows solution
- 17:55Why Argo Workflows chosen
- 21:46Early results and performance
- 23:10Implementation recipe
- 27:47Key takeaways
- 30:15Closing remarks & Q&A
related talks
27mLeveraging Argo Workflows To Move, Migrate and Manage All of Your Workflows
20mDemystifying Argo Workflows: An Architectural Deep Dive
05mArgo Workflows 4, What's New And What's Next
36mAutomating Cloud-native Spark Jobs with Argo Workflows
52mCloud-Native Kubernetes Workflows on AKS with Argo
18mFrom Click To Chaos: Linking Argo Workflows and Backstage for Automate...
27mBuilding a Data Science Platform with Argo Workflows
05mArgo Workflows Update