Large Website Benchmarking

[ayuhito]

To validate the effects of #7, we should have a much much larger benchmarking and staging test suite to verify the performance of huge websites and ingest limits. This should also let us give more predictable estimates of storage used in the docs.

Ideally, we should let this "website" be accessible on the demo website as well. I'm thinking around 50-100M events would be reasonable. Currently, our test suite uses a golden fixture consisting of 1M events.