Story (Required)

As an ArgoCD administrator trying to tune performance of ArgoCD I want to analyze the cache and memory usage of an ArgoCD instance and identify parameters to fine tune the application controller pods.

Background (Required)

In a cluster scoped ArgoCD instance, with 100 namespaces and ~10 applications per namespace, analyze the resources consumed by the instance. There is a known issue where the API server throttles the requests from application controller pod. Check if there are other resource related bottlenecks.

Out of scope

Namespace scoped instances are not covered as part of this story.

Approach (Required)

Analyze how much of memory is used by the application controller pod, redis cache, repo-server.

• Utilize the scripts available in this repo for creating the base setup. https://github.com/akram/argocd-scalability-test/tree/main
• Use prometheus and grafana dashboards for capturing the metrics. (CPU, Memory, Goroutines, Cache size)
• Use Parca (https://www.parca.dev/) to store the profile data and understand where there resources are being consumed.

Dependencies

None

Acceptance Criteria (Mandatory)

• Create a load testing setup for cluster scoped instance
• Capture the resource usage and other metrics using profiling tools and analyze the usage.
• Capture the time delay for application sync.
• Check if there are any application sync lags caused due to

INVEST Checklist

Dependencies identified

Blockers noted and expected delivery timelines set

Design is implementable

Acceptance criteria agreed upon

Story estimated

Legend

Unknown

Verified

Unsatisfied

Done Checklist

• Code is completed, reviewed, documented and checked in
• Unit and integration test automation have been delivered and running cleanly in continuous integration/staging/canary environment
• Continuous Delivery pipeline(s) is able to proceed with new code included
• Customer facing documentation, API docs etc. are produced/updated, reviewed and published
• Acceptance criteria are met