Unleash ArgoCD Observability Superpowers
summary
Intuit runs 44 ArgoCD instances managing ~16,000 applications and uses built‑in Prometheus metrics plus Grafana dashboards with pager links to triage incidents. ArgoCD emits traces via ArgoCD 2.4 to any OpenTelemetry collector (examples shown with Jaeger) and is configurable through argo-cd-cmd-params ConfigMap. Continuous profiling with Parca and Periscope finds real bugs: owner‑list permission checks inflate app-list latency from 40s to 7s (then to 2s after fixes), and unmarshalling huge cluster secrets causes CPU spikes until upstream adds an index.
AI-written summary · about
chapters (10)
- 0:00Intro
- 0:42The Problem
- 2:08What are the Argo CD superpowers?
- 4:13Prometheus Metrics
- 10:33Distributed Tracing
- 19:19What is Continuous Profiling [1][2]
- 21:02Continuous profiling On ByteDance Edge Platform
- 23:18ArgoCD slowness Case study 1: Rbac.enforce
- 24:54ArgoCD slowness Case study 2: secretToCluster
- 27:32Conclusion
related talks
23mArgo at Scale: Navigating Complex Multi-Dimensional Deployments Across H...
18mUnveil the Secret Ingredients for Argo CD in the Enterprise-Scale
15mA Quantitative Study on Argo Scalability
25mSailing Across the Ocean of Argo Ecosystem Like a Pro
22mReal Time Argo Rollouts Analytics Powered by Notification Engine
27mWhy Is It Taking so Long?! Shining the Light on Application..
23mStop Deploying Blind! Using Observability and Argo Rollouts to Light the Way
10mKeynote: Pushing the Limits: Scaling ArgoCD’s Repo-Server To Handle 30...