This is a small node app that profiles a list of sites and corresponding urls for each site gathering the time it takes to complete the request. It is supposed to give a general representation of the speed at which a restual api service is fulfilling requests. The requests are made in a synchronous manner as to only hit one endpoint at a time. This method of request queuing with ensure that each framework has its full CPU Cycle potential to complete the request.
I created it to test the speed of a few different frameworks (Ruby on Rails, CakePHP, NodeJS Express, and Python's Django) in helping to illustrate the difference in response times. Each framework will be sorting through the same list of items and returning the same result to try to keep the scientific controls in a good place.
git clone https://github.com/markhatchell/node-rest-api-profiler.git
cd node-rest-api-profiler
npm install
You will then need to specify the sites and urls you would like to profile.
There is a site.example.js file that demonstrates the object structure for the list of sites and urls to profile.
Make a copy of the sites.example.js file renaming it to sites.js.
Add your list of sites and urls.
You are now ready to run the profiler.
The app is written with ES6 and uses babel to compile back to es5.
To run the app
npm start
npm start will compile the app using babel and then run the built app.
I used JEST and Nock for testing and request mocking.
To run the tests simply run
npm run test
- Display a tables of results in the terminal rather then a list of urls and times
- Display overall comparison of results between frameworks
- Write a React front end for easy of use and display of results with data a visualization library
- Implement CD and publish as a web app.