The Service Delivery team (led by akonarde) will be primarily responsible for Site Reliability Engineering (SRE) of RHODS in production. We need to ensure that RHODS meets all requirements they have for taking ownership of a service in production. This epic tracks all of that work. Work that is not absolutely required for the general availability launch (by July 2021) will not be tracked here.

Work required for limited availability launch (Summit) is tracked in RHODS-265

Expected user experience

1. Purchase RHODS through Marketplace 
2. Request quota for different types of GPU in OCM Resources 
3. Install/upgrade/uninstall Nvidia GPU operator and Node Feature Discovery through OperatorHub 
4. Launch JupyterHub

• Notes:
  • It has not been decided whether the GPU operator will be a separate add-on or part of the platform. 
• Additional Info 
  • RHODS Service Definition
  • RHODS FAQ
  • Violet Intern Project: RHODS and GPU Support