meta: setup continuous benchmarking

[ankush]

1. Get a new bare metal server
2. Do some A/A tests ensure <1% variance
3. Setup self hosted GHA runner
4. Write an action for benchmarking - nightly + dispatch
5. Do some A/A testing with GHA setup.
6. Add custom metadata to pyperf's JSON output.
7. Evaluate bencher v. custom insights based solution
8. Establish a baseline with this commit frappe/frappe@7c06c84
9. Run benchmarks nightly
10. Upload data somewhere
11. Ensure concurrency of 1
12. Dashboards, pretty charts
13. Alerts

[epompeii]

@ankush you may want to check out Bencher: https://github.com/bencherdev/bencher

Bencher supports 2, 3, 5, & 6 out of the box, for free for public projects. Points 1 & 4 are also available. I'm the maintainer of Bencher, so I would be more than happy to help answer any questions.