The POC is a group of four microservices using event sourcing to communicate what is happening in each context.
The main js files are:
store-api.js: simulation of a API, receiving order info and writing it in a t_orders topic.finance.js: simulation of a service that consumes t_orders, do something related to finance rules, and then write new info in a t_receipts topic.logistics.js: simulation of a logistics service consuming t_receipts topic, doing some delivery tasks and writing a new message in t_deliveries topic.mailer.js: this communication service consumes two different topics, t_receipts and t_deliveries, and reacts to these two events producing a new message it t_emails topic for each message handled.
I did this POC using Zookeeper and Kafka running on my machine. It's possible do it following the steps 1 and 2 in this tutorial: https://kafka.apache.org/quickstart
To wrap up, after Kafka download and extraction, run:
- Start Zookeeper coordinator:
# bin/zookeeper-server-start.sh config/zookeeper.properties - Start Kafka server:
# bin/kafka-server-start.sh config/server.properties
It's easy deploy Jaeger for study using Docker:
# 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 14250:14250 \
-p 9411:9411 \
jaegertracing/all-in-one:1.21
And access de Jaeger User Interface at http://localhost:16686/search.
- Open five terminals and execute:
- node finance.js
- node logistics.js
- node mailer.js
The three apps will start and subscribe their Kafka topics.
- Run store-api.js to start the real thing.
Each time store-api is executed, it will be produced a message in t_orders, which will unfold the processing on the others services.
There is a quering interface (in this case, just the service was chosen):
It's possible visualize details of transactions just clicking at the item:
This is an experiment using jaeger-client package to achieve Distributed Tracing in a microservices architecture that uses Kafka to stablish asynchonous communication. The way that jaeger and kafka components are initialized here are not recomended to production environment.