Write additonal asv "airspeed velocity" benchmarks for conda; set up scheduled benchmark runs.

[dholth]

Checklist

• I added a descriptive title
• I searched open requests and couldn't find a duplicate

What is the idea?

https://github.com/conda/conda-benchmarks uses airspeed velocity or "asv" to run benchmarks against conda. The asv benchmarker can run tests against multiple revisions of a program to track and graph performance over time. The repository has a recently updated set of benchmarks especially related to parallel downloads and the jlap feature.

Write additional benchmarks; set up automated benchmark runners (in GitHub Actions would be great) or recruit people to run them on various hardware.

Why is this needed?

By tracking conda's performace over time we can improve that performance or avoid regressions.