Helm Values to Configure Argo CD Cluster Cache Resync Durations

[yaswanthvarma]

Argo CD version: 3.1.8
Environment: Rancher-managed multi-cluster setup (Rancher + AKS)

Summary
We are operating Argo CD in a large multi-cluster environment and have observed unexpectedly high API traffic from Argo CD to each downstream Kubernetes cluster. Even after increasing the application reconciliation period (from 3 minutes to 4 hours), the Kubernetes API servers still receive large volumes of LIST calls from Argo CD.

During audit log analysis, we observed that ArgoCD performs GET calls against all Kubernetes API groups and CRDs every 10 minutes.
These calls originate from the argocd-application-controller user-agent and match the intervals defined in ArgoCD’s internal cluster cache logic.

In the source code (v3.1.8), these values are hard-coded:
https://github.com/argoproj/argo-cd/blob/v3.1.8/controller/cache/cache.go

      // GitOps engine cluster cache tuning options
     var (
           // Full cluster cache refresh
           clusterCacheResyncDuration = 12 * time.Hour
          // Forces a relist of all watches (API groups, CRDs)
          clusterCacheWatchResyncDuration = 10 * time.Minute
          )

The clusterCacheWatchResyncDuration (10 minutes) appears to drive repeated LIST operations against:

• /api/* (core Kubernetes resources)
• /apis/* (CRDs & non-core API resources)

This causes ArgoCD to send large-volume, cluster-wide LIST requests every 10 minutes, across all API groups—even for CRDs used by applications.
We couldn't find a Helm chart value or environment variable that allows users to configure this 10-minute resync interval.

Please advise if there is a way we could influence this setting.
This would allow deployments to significantly reduce load on the Kubernetes API server while still maintaining Argo CD functionality.