Investigate whether or not parallelism is beneficial towards the ETL process

[paul-tqh-nguyen]

We need to profile our ETL process to see how well it's performing and what's slowing it down.

If most of the time is spent processing the data, we can speed this up via naive parallelism.

One benefit fo the slowness is that it'll help prevent throttling.

[paul-tqh-nguyen]

Nope, we get throttled pretty quickly.

That's how I learned about https://arxiv.org/denied.html