This code is running at envoy-playground.apoxy.dev.

It is based on the nginx-playground project.

This project is maintained by Apoxy. If you find an issue feel free to either open a pull request with the fix or create an issue.

If you fork this project again, jvns has noted a few things that are specific to this deployment of this code, that you'll want to remove or change if you make significant changes:

• the header
• the analytics (grep for posthog)
• the fly.toml
• the FAQ

Might have security vulnerabilities, it gives the user a lot of access, I personally would only run this software on a machine that I was comfortable with potentially being compromised. I haven't had any issues that I know of yet though.

I run the backend and the frontend on separate servers so that if the backend ever did get compromised, it wouldn't affect the frontend site. You can just run the frontend on GitHub pages or something.

Developing the frontend is straightforward:

cd static
python3 -m http.server 8084 # or serve a webserver any other way you want

It has the URL of the backend hardcoded so you I work on it without having to run the backend.

To generate the Tailwind CSS: (with the tailwind standalone CLI)

tailwindcss-macos-arm64 --content  'static/*.html,static/*.js' -o static/css/tailwind-classes.css

Developing the backend is a bit of a mess. It only works on Linux (because it depends on bubblewrap), and it's a pain to develop because it requires nginx to be installed in specific directories and it has to run as root.

Here are setup instructions that worked for me on a fresh Ubuntu 22.10 install:

sudo apt-get install golang nginx bubblewrap
# install go-httpbin
go install github.com/mccutchen/go-httpbin/v2/cmd/go-httpbin@latest
# just putting go-httpbin in your PATH doesn't work
cp /the/path/to/go-httpbin /usr/bin 
# there has to be an nginx user
useradd -s /bin/false nginx 
# start the server
bash scripts/run-local.sh

You can test that it's working with httpie like this:

http post localhost:8080/run nginx_config=@examples/basic.conf command="http get localhost"

There's also a watch-local.sh that watches your changes with entr.

It should in theory be possible to modify the way bubblewrap is invoked to run without being root (and have a less scary development experience), but I couldn't figure out how to do it. If you figure out how to make the experience of working with bubblewrap easier I'd love to hear about it.

Also you could definitely modify this so that it works with a separate nginx container image and doesn't require you to have nginx installed in your main filesystem.