Closes#10600

I'm trying to understand why these tests flake sometimes, like here:https://github.com/coder/coder/actions/runs/6805426561/job/18504925666#step:5:428

Here's what I have so far:

• The tests failed very close to midnight:insights_internal_test.go:162: now: 2023-11-08 23:59:56.781794 +0000 UTC m=+7.293516543
  • I noticed the monotonic clock reading,and upon further reading about it I think it's reasonable to strip this when dealing with parsed times since we want our operations to match the presentation time. Nvm, time.Parse already has this stripped so that won't do anything...
  • Some tests fail with time not being an even hour, which I believe could be caused by a difference in wall vs monotonic clock. We now think this is because we create a newtime.Now() inside the function, and with processing lag this can be significant enough to change the time returned.
• I've added more logging aroundLocation and other times in the tests, as well as more verbose error responses so we know what times it thinks it got. My hunch is that all of these tests fail when it runs on a CI machine that has different time settings than we are used to running in.

With the logs we hopefully have a better shot next time this happens.

Update:

• After speaking with@mafredri we now pass thenow time into the function. This is so if we are processing the test very slowly (like in CI) the drift can cause flakes when on a time barrier (end of hour, end of day, etc).