Also the following metadata:

• test environment (Evergreen, Github Actions, Other)
• operating system (Windows, Mac, Ubuntu, RHEL)
• branch/pr/commit/patch
• requestor or whether it was scheduled

We'll have to find a way to uniquely identify each test or hook. These will change over time, but that's OK.

Most important is execution time and whether the tests passed or failed. It might be tempting to store the coverage % as well, but the issue to add coverage will likely play out in parallel with this one, so maybe do that in a second pass.