Drop Vulture Loki dependency

[dgzlopes]

What this PR does:
This PR refactors Tempo Vulture to remove the actual Loki dep.

Now, instead of searching for traces in Loki we:

• Generate a new trace (with some random data/spans).
• Push the trace to Tempo (in batches).
• Query a "pushed" trace randomly.

Which issue(s) this PR fixes:
Fixes #374

Checklist

• Tests updated
• Documentation added
• CHANGELOG.md updated - the order of entries should be [CHANGE], [FEATURE], [ENHANCEMENT], [BUGFIX]

[joe-elliott]

Great progress!

I tried to comment on what I was looking for w/r to the random # seeding. If that's unclear please reach out.

[joe-elliott]

one nit, then i'll merge.

[joe-elliott]

this is really close. one last comment, but i can be persuaded otherwise.

[joe-elliott]

i'd like to drop this slot idea and just generate the directly on the timestamps. i think this will make things easier to replicate outside the vulture.

i also think we should split the read/write tickers. maybe we only want to generate a trace every 30 seconds, but execute a query every 5 seconds?

[dgzlopes]

i'd like to drop this slot idea and just generate the directly on the timestamps. i think this will make things easier to replicate outside the vulture.

I removed the slot thing but from my POV the result was not great... Maybe I misunderstood something...

But, I think your interval approach is great, so I refactored the code to replicate that behavior. Now it's much cleaner and works flawlessly. Hopefully, it makes some sense!

i also think we should split the read/write tickers. maybe we only want to generate a trace every 30 seconds, but execute a query every 5 seconds?

Done :)