Benchmarks


This is the final output of an internal benchmark run in github action on 4/16/2024, 12:53:29 AM. It consists of several simple applications for specific purpose. Each is then accessed by the OHA within 10s. The results are then sorted by the fastest.

You can find the benchmark script in this code: run.ts

Benchmark results

module rps % oha cmd
deno 57154 100% oha -j --no-tui -z 10s http://localhost:8000
ctx_string 50771 89% oha -j --no-tui -z 10s http://localhost:8000
group 50681 89% oha -j --no-tui -z 10s http://localhost:8000/api/user
string_response 49985 87% oha -j --no-tui -z 10s http://localhost:8000
raw_string 49719 87% oha -j --no-tui -z 10s http://localhost:8000
raw_json 49588 87% oha -j --no-tui -z 10s http://localhost:8000
ctx_json 49540 87% oha -j --no-tui -z 10s http://localhost:8000
params_query 47844 84% oha -j --no-tui -z 10s http://localhost:8000/agus?title=lead
ctx_jsx 46955 82% oha -j --no-tui -z 10s http://localhost:8000
app_middleware 42495 74% oha -j --no-tui -z 10s http://localhost:8000
deno_sqlite 34467 60% oha -j --no-tui -z 10s http://localhost:8000
server_rendering 26061 46% oha -j --no-tui -z 10s http://localhost:8000
deno_kv 20684 36% oha -j --no-tui -z 10s http://localhost:8000/user?name=john
markdown_middleware 19697 34% oha -j --no-tui -z 10s http://localhost:8000/blog/hello
deno_redis 15394 27% oha -j --no-tui -z 10s http://localhost:8000
oauth 14524 25% oha -j --no-tui -z 10s http://localhost:8000
deno_mongo 14301 25% oha -j --no-tui -z 10s http://localhost:8000
deno_postgres 8299 15% oha -j --no-tui -z 10s http://localhost:8000
deno_mysql 7737 14% oha -j --no-tui -z 10s http://localhost:8000
route_middleware 4776 8% oha -j --no-tui -z 10s http://localhost:8000
static_file_string 1935 3% oha -j --no-tui -z 10s http://localhost:8000/static/tailwind.css
static_file_image 1769 3% oha -j --no-tui -z 10s http://localhost:8000/static/favicon.ico
Fastro Framework is free and open source. Powered by Deno Deploy. Requests per second (RPS) for each use case is monitored daily through internal benchmarks.