Stress Server

Source code on GitHub | Image on DockerHub

This is a docker container with the stress command line installed, and
a webserver that can run stress calls with options.

It was designed to help in testing and debugging auto scaling configurations.

Usage

Start the server:

$ docker run --rm -it -p 3000:3000 dannyben/stress-server

Use one of the endpoints below to run the stress command on the server.

Notes

• The server streams its output back to your client. It is recommended you invoke your calls with a streaming option (for example, --no-buffer in curl).
• It is recommended to run all your calls with the timeout option set, otherwise, you might need to restart the server in order to stop the stress process.
• You should not run this server in production environments.

Server Endpoints

GET /stress

This is the primary endpoint. It accepts all the arguments accepted by the stress command line interface, as query string parameters. You may use either the long or short form.

For example, if you wish to perform this command:

$ stress -c 1 -m 1 --verbose -t 10

Use this URL:

$ curl --no-buffer "http://localhost:3000/stress?c=1&m=1&verbose=1&t=10"

Flags with a hyphen, like --dry-run may be expressed with either a hyphen (dry-run=1) or an underscore (dry_run=1)

GET /help

Show the output of stress --help:

$ curl http://localhost:3000/help

GET /raw/:options

This endpoint runs the stress command using options in the same syntax as you would normally provide them in the command line. When using curl or other HTTP client that does not like spaces in the URL, simply replace the spaces with a + sign.

For example, if you wish to perform this command:

$ stress -c 1 -m 1 --verbose -t 10

Use this URL:

$ curl --no-buffer "http://localhost:3000/raw/-c+1+-m+1+--verbose+-t+10"

Development

To run the server in development:

$ bundle exec rackup -s thin -p 3000 -o 0.0.0.0

Or, if you have opcode:

$ op serve

To run the dockerized version of the server:

$ docker-compose up web