Land initial groundwork for better self-hosted testing infrastructure

[hubertdeng123]

We should have a good idea of where we want testing for self-hosted to be in our pipeline, and outline a process for devs to write their own tests for self-hosted. The first step here would be to rewrite our entire testing suite in a framework that is more friendly. If we ourselves struggle to write tests here, how can we expect others to willingly shoulder this burden with us?

Tasks

Preview Give feedback

1. rewrite current e2e tests to a more friendly testing framework
2. measure flakiness of self-hosted e2e tests and ensure it is under <10%
3. lay out new groundwork to re-add testing in the deployment pipeline