nilportugues / jaeger-nodejs-example

An example of Jaeger instrumentation for Node.js micro-services

Using Jaeger with your Node.js micro-services

My motivation for creating this repository 😺

So you are into micro-services and you already found about OpenTracing and you want to use that fancy JaegerUI to see what is really going on between your micro-services. You already noticed that there are not many examples out there about jaeger and node.js. I hear you! This is why I created this repository, I want to help you understand how multi-micro-service tracing works with Jaeger.

You will learn how your tracers in your micro-services inject 💉 and extract 👀 headers.

Requirements 📝

• Node.js Active LTS recommended
• PM2 To start multiple node.js instances easily
• Jaeger All in One Docker Image To setup a jeager collector in your local environment

How to setup all of these and make it work ✨

1. Install prerequisites

   • Make sure you have installed Node.js and Docker

   • Install PM2 globally

     npm i -g pm2

   • Run Jaeger All in One Docker Image

     docker run -d --name jaeger \
     -e COLLECTOR_ZIPKIN_HTTP_PORT=9411 \
     -p 5775:5775/udp \
     -p 6831:6831/udp \
     -p 6832:6832/udp \
     -p 5778:5778 \
     -p 16686:16686 \
     -p 14268:14268 \
     -p 9411:9411 \
     jaegertracing/all-in-one:1.8

2. Run the example project

   • Clone this repository

     git clone [email protected]:algun/jaeger-nodejs-example.git

   • Change your directory into the project

     cd jaeger-nodejs-example

   • Install packages

     npm i

   • Use PM2 to start up all your services

     pm2 start

   • Use PM2 to watch project logs

     pm2 logs

3. Visit your initial endpoint to create traces between all services

   • http://localhost:3000

4. Visit your local JaegerUI to enjoy your micro-service trace visualizations

   • http://localhost:16686/

5. Enjoy these views

Understanding how it works ✨

Here are some key points to understand how this project instruments these services with jaeger:

• Check ecosystem.config.js file to see how service names and ports are defined
• Pay attentions to tracer.inject() and tracer.extract() commands to understand how the spans flow into another service via http headers.
• Notice that init-service has only tracer.inject() since it is the firelight of our multi-micro-service calls.
• Notice that service4 and service5 have only tracer.extract() since they are the end of the line for our micro-service structure in this example.
• Notice that service4 still injects its headers with tracer.inject() but it makes no sense since it will not create request any further within this project.

Feel free to comment / give feedback ❤️

This is my humble attempt to help people out who is struggling to understand tracing concept between micro-services.