Here's you'll find the sample app used to show how midas-cache works, and its benchmark. The README in each project contains set up instructions.
Running the benchmark produces the following result:
hyperfine --warmup 1 'oha --no-tui -n 500 http://localhost:3000/records'
'oha --no-tui -n 500 http://localhost:3001/records'
'oha --no-tui -n 500 http://localhost:3003/records'
'oha --no-tui -n 500 http://localhost:3004/records'
Result:
Summary
oha --no-tui -n 500 http://localhost:3003/records ran
1.39 ± 1.90 times faster than oha --no-tui -n 500 http://localhost:3004/records
3.76 ± 0.38 times faster than oha --no-tui -n 500 http://localhost:3000/records
32.73 ± 3.42 times faster than oha --no-tui -n 500 http://localhost:3001/records