GoTo Financial’s Story: Towards 10k ArgoCD Apps to Support...

on YouTube ↗30m

" summary

GoTo Financial runs a centralized ArgoCD across ~40 Istio-enabled clusters to manage ~10k apps and ~5k repos and hits scale pains: UI loads take minutes, repo-server OOMs, and controllers saturate CPU. Concrete fixes include gzip compression (5x faster UI, 7x less transfer), removing CPU limits on app-controller to avoid throttling, scaling repo-server with HPA plus parallelism_limit, raising repo cache expiry to cut git fetches, manual shard allocation for even CPU, and ignore-resource-updates to halve controller CPU.

AI-written summary · about

" chapters (19)

1. 0:00GoTo Financial’s ArgoCD journey
2. 1:05ArgoCD setup and scale challenges
3. 3:55ArgoCD usage patterns
4. 6:46ArgoCD architecture issues
5. 8:15Optimizing ArgoCD server UI
6. 9:22UI application filtering
7. 10:03Kubernetes CPU limits tuning
8. 11:30Repo server scaling OOM
9. 12:56Timeout errors in manifests
10. 13:38Reducing Git API calls
11. 14:44Monorepo challenges
12. 16:47Reconciliation improvements
13. 17:27App controller tuning
14. 18:29App controller sharding
15. 19:56Ignoring high-turnover objects
16. 22:07API client library fixes
17. 23:32Decentralized ArgoCD models
18. 25:21Hybrid model improvements
19. 27:09Q&A