deployment-controller

This project is the top layer to my webserver. It features a reverse proxy that routes all the request via an auth request to the expected projects, a backend that handles the auth requests and a small UI for a small overview and a login screen.

features

• n projects behind one domain, with subdomains or locations
• optionally require login to access project
• actually secure login (? :))
• store who accessed the pages
• visualize who accessed the pages
• logging, that makes performance and quality reviews easy
• send informative mails to the admin and the users

tech stack

• nginx
• postgres
• go + gin
• sveltekit
• elastic, kibana, logstash

dev

When developing, it helps to have a reverse proxy setup, so that everything can be tested. To make this process easy, I have created the dev-proxy dir.

proxy

docker compose up -d -f ./dev-env/docker-compose.yml

subdomains
To use subdomains in localhost (on mac) I had to modify the /etc/hosts file. I added lines like this:

127.0.0.1 michu-tech-dev.net
127.0.0.1 host.michu-tech-dev.net
127.0.0.1 host.backend.michu-tech-dev.net
127.0.0.1 teachu.michu-tech-dev.net
127.0.0.1 room-automation.michu-tech-dev.net
127.0.0.1 kibana.michu-tech-dev.net

k6

k6 is a test application that helps with testing truly parallel requests.
This needs to be installed on the developers machine, otherwise the tests in ./backend-k6-test/script.js won't run.

db

The file ./db/init.sql creates the DB schema.
The file ./dev-proxy/test-data.sql inserts some test pages, so that you can test.
A host page must exist. (even the host page has an auth request)

elastic

Using Elasicsearch, Kibana, Metricbeat and so on to keep an overview of my system.

Here is a list of things I plan on doing with the Elastic Stack

• Nginx Metrics and Logs
• Postgres Metrics
• Docker Metrics
• Docker Container logs
  • since I have a lot of projects that just run in a docker container, it will be very usefull to see the all the logs without having to SSH onto the server.
• Backend with APM
  • Requests should be logged
  • Errors should be reported
  • (maybe also log DB Queries, Postgres integration might be enough)
• Agent on Raspberry PI
  • Get metrics and maybe logs from my raspberry pi
• Helpful Kibana dashboards
• Security or error alerts

backend

The backend can usually be started in the IDE or with go run main.go.
In production the app will run in a docker container. To test the container run the following.
The docker container uses the config-docker.yml config!

docker build --tag deployment_controller_dev_backend ./backend

docker run -p 8080:8080 --name dp_crtl_be deployment_controller_dev_backend

frontend

The frontend can be started and built with the following command.
Make sure that a .env file exists with the PUBLIC_BACKEND_URL config.

# run the dev builds
npm run dev

# compile for production
npm run build