shinycannon

A load generation tool for Shiny applications. Part of https://rstudio.github.io/shinyloadtest

Installation

For installation instructions, please see the shinyloadtest documentation.

Development

shinycannon is written in Kotlin, a JVM language. Consequently, shinycannon is able to run on whatever platforms JVM is.

However, to ease installation on various platforms, we produce package installer files using fpm in addition to a .jar file.

Building

Building packages for all platforms is best accomplished with Docker. We'll use the Dockerfile included in this repository.

First, build the Docker image with a command like the following:

docker build -t shinycannon-build .

Executable

To build the .jar file:

sudo docker run -it --rm -v $PWD:/root -w /root shinycannon-build make maven

Note: you may or may not need sudo, depending on how you installed Docker.

A file, target/shinycannon-1.1.0-jar-with-dependencies.jar will be created.

Installers

To build the .jar file and all packages (rpm/deb/sh):

sudo docker run -it --rm -v $PWD:/root -w /root shinycannon-build make clean maven packages

Note: you may or may not need sudo, depending on how you installed Docker.

All installer files will be stored in the ./out folder

Running

You can run shinycannon with something like:

java -jar \
  ./out/shinycannon-1.1.0-45731f0.jar \
  recording.log \
  http://example.com/your/app

If you are running against an app on RStudio Connect, an RStudio Connect API Key is the easiest way to handle authentication. Please visit Load Testing Authenticated Apps for examples on how to safely set a sensitive environment variable such as SHINYCANNON_CONNECT_API_KEY.

Note: that if the recording was done with an RStudio Connect API key, playback MUST be done with an RStudio Connect API key. Similarly, if a recording does NOT use an API key, playback must NOT use an API key.

Releasing

In the past, release installers were built privately on internal RStudio infrastructure. Release artifacts were deployed to S3. Now, release artifacts are built with a GitHub Actions workflow and uploaded as GitHub release assets.

Before releasing, first update pom.xml with the desired version. Then, to create a release and build corresponding release artifacts, create and push a tag prefixed with v such as v1.2.3. On push, the shinycannon .jar and a set of platform-specific installers will be created and uploaded.

Releases are available under the "Releases" tab of the project GitHub page.

After release artifacts are created, you should update the shinyloadtest documentation to point to the correct URLs.

License

MIT

rstudio/shinycannon