A Quantitative Study on Argo CD Scalability

Explore a quantitative study on Argo CD scalability in this 15-minute conference talk presented by Andrew Anderson and Jun Duan from IBM at a CNCF event. Delve into the research focused on determining the maximum number of Argo CD Applications that can be supported while maintaining reasonable performance. Examine the key performance metric of syncing delay between changes in upstream git repositories and resources in downstream Kubernetes clusters. Discover how CPU becomes the primary bottleneck when dealing with thousands of Argo CD Applications. Learn about the multi-faceted approach to addressing this challenge, including reducing computational requirements, increasing computing power, and adjusting expectations for syncing delays. Gain insights from the presenters' experiments, which successfully synchronized 10,000 Argo CD applications in under 40 minutes.

A Quantitative Study on Argo Scalability - Andrew Anderson & Jun Duan, IBM