In an unrelated MR, a conversation came up about whether it would be faster if we did X instead of Y.  This sort of conversation is difficult to resolve without understanding the base performance of our products.  We should create some basic benchmarks that allow us to run `make bench` and see some results.  Ideally, this should over API, logic, and DB related functions using golangs built-in benchmarking tool.  Presumably, we can have Unit-level benchmarks and Integration-level benchmarks, similar to how we have unit tests and integration tests today.