URL-Shortener

This is an implementation of a URL-Shortener in Go with Redis as the in-memory database, using a fixed window counter algorithm, which is a standardized rate limiting algorithm. The goal of this project is to test different rate-limiting algorithms on client IPs and how they affect hit rates, as well as to provide real-time analytics for URLs. The current version only contains the APIs with limited performance testing, further versions will contain extensive stress testing alongside an interactive frontend for users. All the components are dockerised. The containers are then monitored using Prometheus and Grafana.

Tech-Stack

• Languages: Go, YAML, JavaScript
• Databases: Redis: used RedisInsight to manage instances
• Deployment Tools: Docker: used Docker-Compose to manage multi-container applications
• API Testing Tools: Postman, Apache JMeter
• Monitoring Tools: Prometheus and Grafana

Guide: How to deploy locally

1. Clone the GitHub repository using the following command to your local machine.

   git clone https://github.com/sankalp-12/URL-Shortener
   

2. Make sure you have installed Docker and Docker-Compose on your machine. Now, change the working directory to the newly cloned repository.

   cd URL-Shortener
   

3. Now, to start all the containers, run the docker-compose.yml file.

   sudo docker-compose build && sudo docker-compose up
   

   Alternatively, a Makefile is provided to restart the process in case of any errors. If you face any errors during the executions of the YAML config file, ensure that all the ports required are free on your machine.

4. After execution, run the command: docker ps to ensure that all the containers are up and running.

5. Once verified, access the port(9090 by default) assigned to the Prometheus instance. Navigate to the Status option in the panel, and click on the Targets option in the drop-down menu. If you see 3 UP instances, then Prometheus is successfully scraping the application.

6. Finally, access the port(3031 by default) assigned to the Grafana instance and start making your dashboards to monitor your application metrics and send alerts. Examples of possible dashboards you can create are given below.

7. Note that to get metrics for a locally deployed application, you will have to hit the APIs yourself. Postman and JMeter were used for performance testing of the application to capture the above metrics. For Postman specifically, you can use pre-test scripts written in vanilla JavaScript to automate performance tests. However, if you want to automate stress tests in particular, JMeter is recommended as it allows you to engage in IP Spoofing to test your application from different IPs, in case you have a single machine for load testing.