Test GitLab Runner Manager Pod performance and provide best practices

Overview

Test GitLab Runner Manager Pod performance and provide best practices

this issue is a follow-up of the MR !167209 (merged)

In the linked MR, the tests focused on job Pods and highlighted best practices to ensure efficient workload execution on the Kubernetes cluster environment at hand.

This issue, on the other hand, aims to address best practices and recommendations specifically for the Pod running the Runner Manager. This Pod is likely to handle:

• Log processing for all jobs being executed
• Caching management, both locally (e.g., persistent volumes) and with supported cloud providers
• K8S API requests, including retries in case of network failures
• GitLab API requests for managing concurrency, etc.

These are the key points I had in mind when creating the issue.

Testing Goals

• Provide best practices for GitLab Runner Manager Pod
• Provide tools to help customers take right decision when deploying GitLab Runner Manager on a k8s/OpenShift cluster

Customer requirements

• Customers need to know how to determine when the runner manager pod is overloaded and the best practices to the performance of the runner manager pod.