Resource requests for default instance are excessive for small use cases

[davgordo]

Describe the bug
When using OLM to provision the OpenShift GitOps operator, a default ArgoCD instance is deployed to the openshift-gitops namespace. Recently, default resource requests (and limits) were introduced to this default instance. The chosen values seem to exceed what is necessary for essential ArgoCD functionality.

Perhaps resource requests should be informed by the intended use case and therefore not imposed by the default template to accommodate the broadest range of use cases. Requests can be tuned to specific use cases at a later point in time. It is potentially harder to backtrack if a small environment is oversubscribed with resource requests.

To Reproduce
Steps to reproduce the behavior:

1. Run OpenShift Local on workstation with average resources (e.g. ~4 vCPU, 12-24Gi memory)
2. Use OLM to install OpenShift GitOps (including default ArgoCD instance)
3. Note requests and events for OpenShift GitOps pods (in my experience all CPU was consumed, and some components could not be scheduled)
4. Apply smaller resource requests with an ArgoCD custom resource (in my experience, it had no effect because all CPU budget is already spent with a queue of large pods in pending status)
5. Manually interfere by deleting large pending pods (in my experience this resolves the problem, and I could operate ArgoCD with a small fraction of resources compared to the originally requested resources from the default instance template)

Expected behavior
The OpenShift GitOps user experience in small cluster environments should not request more resources than necessary for small use cases.