Server-Side Diff: Future-Proofing ArgoCD's Comparison Engine
summary
Argo CD moves diff computation to the Kubernetes API using Server-Side Apply dry-run, generating predicted live state that respects CRD schemas, defaulted fields and admission webhooks (mutation hooks omitted by default, opt-in). It tracks ownership via manageFields ownership, removes reliance on last-applied annotation, and scales to Intuit-class environments (300+ clusters, 3,000+ services, 50k namespaces). Expect about a ~3s / ~30% per-app CLI cost in benchmarks; mitigations include aggressive caching, skipping server-side diff on resource creation, and gradual rollout with an annotation or controller-wide flag.
AI-written summary · about
chapters (14)
- 0:00Server-side diff intro
- 1:04Why server-side diff matters
- 3:33Client-side diff problems
- 9:35How server-side diff works
- 11:46Advantages of server-side diff
- 13:12Catching issues before syncing
- 15:19Admission webhooks in server-side diff
- 17:13Field ownership tracking
- 18:36Better CRD support
- 20:20Drawbacks of server-side diff
- 21:47Performance impact
- 23:12Implementation lessons
- 24:15Enabling server-side diff
- 24:36Key takeaways
related talks
29mHow to Preview and Diff Your Argo CD Deployments
05mProject Lightning Talk: Argo CD at Scale: 5 Features You Should Not Miss!
28mArgo CD Maintainer's Panel
18mUnveil the Secret Ingredients for Argo CD in the Enterprise-Scale
28mClick-Free Environment Promotion Driven by ApplicationSets Progre...
06mProject Lightning Talk: Argo CD at Scale: 5 Features You Should Not Miss!
25mSailing Across the Ocean of Argo Ecosystem Like a Pro
34mInfrastructure at Scale Using ArgoCD