Update tests to improve CI logging to better diagnose timeout

[allenporter]

Remove tests that often timeout on CI. This will result in lower test coverage, but probably not worth it given the flakes.

[Lash-L]

Looks to still flake unfortunately

Any easy way in purest to just retry?

Retry each test twice or something? I get that that wouldn't be perfect either though

[allenporter]

I reverted removing the tests and added explicit shutdown calls at the end of each test. I also turned up logging on CI with thread names and timestamps.

Funnily, its not timing out as often now, not sure if its because logging triggers I/O which causes the race to behave differently or something else.