We use Github action to run Service Mesh Performance for different service meshes benchmarking, and there are two implementations:
- Configurable Benchmarking Test: Users can provide their own configuration, choose different service meshes and load generators.
- Scheduled Benchmarking Test: The matrix of different service meshes, load generators and configurations will be run periodically.
At this stage, all the tests run on the GitHub-hosted runners, which restrict greatly for tests running. Therefore, we need to extend them to CNCF Community Infrastructure Lab (CIL) to increase the flexibility.
I have drafted a design of “Extending SMP to CNCF Community Infrastructure Lab” : https://docs.google.com/document/d/16b-cWkyhDwumK8afd0YiQtZUe2zc5TfQ_9g7dQ9q8jQ/edit#heading=h.30j0zll
Please feel free to leave comments, thanks!