## Overview Account teams are spending a lot of time creating runner fleet customer's guidelines for the setup and scaling of a Runner Fleet. As there are many dimensions to consider when planning a Runner Fleet for CI at the enterprise scale, we propose creating a library of case studies that technical account managers and customers can reference when considering how best to configure Runners for their specific needs. ## Customer verbatims / requirements - [ ] What are the optimal major Cloud (i.e. AWS & GCP) scale settings? - [ ] How can we achieve cost chargebacks for runner usage to CI users? - [ ] Do you have an Infrastructure as Code deployment or reference architecture for a Runner Fleet on a target cloud platform? - [ ] What are best practices for monitoring runner managers in Kubernetes? - [ ] Do we have any guidelines for runner configuration in terms of quantity, concurrency config for a specified number of projects and an expected build frequency? - [ ] Should you have centralized team handling runners or dev teams handling themselves? - [ ] Do you have any recommended levels for how long, or how many jobs, a runner should run before being restarted ## Proposal - Create a library of best practices for configuring and scaling a GitLab Runner Fleet. ## Links [Current documentation on planning and operating a runner fleet](https://docs.gitlab.com/runner/fleet_scaling/)