Kafka Streamer

A microservice that uses Kafka to listens to an event involving order creation and then insert the product orders to a PostgreSQL database.

Table of Contents

1. About The Project
2. Built With
3. Getting Started
   • Prerequisites
   • Running the project

ℹ️ About The Project

Kafka Streamer is a microservice that uses Kafka to listens to an event involving order creation and then insert the product orders to a Postgres database. The database operations are carried out using a DB sink connector which reads product_orders topic and writes the data to a table called orders in Postgres DB.

The overall goal of the assignment is to build a system that accepts product orders, stores them in a PostgreSQL database, forwards them to city-specific topics in Kafka, and sends an email to customers for high-priority orders.

🛠️ Built With

Following technologies and libraries are used for the development of this project.

• Python
• Flask
• Kafka
• PostgreSQL
• Docker
• Docker Compose
• Shell/CLI

📌 Getting Started

To setup the project locally follow the steps below

💻 Prerequisites

• Docker
• Docker Compose
• Terminal

🤖 Running the project.

1. Fork and clone the project to your local system
2. cd into the project and run

docker-compose build --no-cache
docker-compose up

3. Then we have to create a table in postgre by running the following commands:

docker ps -a

<!-- copy the container id of db container -->

docker exec -it <container_id> /bin/bash

psql -U pritishsamal -p 5432 -h localhost -d order

<!-- Now run the following query: -->

CREATE TABLE orders (
  order_id SERIAL PRIMARY KEY,
  name VARCHAR(255) NOT NULL,
  email VARCHAR(255) NOT NULL,
  street VARCHAR(255) NOT NULL,
  city VARCHAR(255) NOT NULL,
  state VARCHAR(255) NOT NULL,
  postal_code VARCHAR(10) NOT NULL,
  product_name VARCHAR(255) NOT NULL,
  quantity INTEGER NOT NULL,
  order_date DATE NOT NULL,
  priority VARCHAR(255) NOT NULL
);

Then if you run \dt, you'll be able to see a table named orders

4. Now, let's exec into kafka container & create a topic named product_orders. Run the following commands:

docker ps -a

<!-- copy the container id of kafka container -->

docker exec -it <container_id> /bin/bash

cd /opt/bitnami/kafka/bin

./kafka-topics.sh --create --topic product_orders --bootstrap-server localhost:9092 --partitions 1 --replication-factor 1

<!-- The following command will list the topics -->

./kafka-topics.sh --list --bootstrap-server localhost:9092

<!-- To run the consumer and see what events it is listening to in real-time, run -->
kafka-console-consumer.sh --topic "product_orders" --from-beginning --bootstrap-server localhost:9092

<!-- To run the producer for testing purposes and publish dummy orders, run -->
kafka-console-producer.sh --topic "product_orders" --bootstrap-server localhost:9092

5. Now we have the server, broker as well as the database up and running.

6. If visit 127.0.0.1:5000, you'll find the flask server running in this port. Now open postman and send a POST request to 127.0.0.1:5000/orders with the following payload:

{
	"order_id": 90123,
	"customer": {
		"name": "James Clark",
		"email": "james@gmail.com",
		"address": {
			"street": "Avenue Street",
			"city": "Illinois",
			"state": "Chicago",
			"postal_code": "768987"
		}
	},
	"product_name": "Protein Powder",
	"quantity": 5,
	"order_date": "2023-06-09",
	"priority": "high"
}

7. When we send a POST request to /orders endpoint, the following set of events are carried out:

• the json is extracted from request body
• then, it will call publish_to_kafka_topic function to initialize a producer and send data to the kafka topic product_orders
• Since the consumer is always running in a separate thread, it'll be able to listen whenever any new order is added to the product_orders topic via the consume_and_send_emails function
• In this function, first the consumer is initialized and then we iterate through the consumer variable.
• We then check the priority of the order. If it is high then we call the send_email function and then call the save_order_to_postgres(order) to write to db else directly write to db if priority is medium or low.
• Once it is added to database via consumer, we then extract the city from the order object and invoke create_topic_if_not_exists function. This will create a new topic with the name of the city if not already present.
• Once the respective city topic is created, publish_to_kafka_topic function will be called which will publish the order to it's respective city topic.
• At last, the api will return a success message

8. If we send a GET request to /orders endpoint, it will simply establish connection with the db and then get all the orders and send it in the form of a json.

📉 Relevant Screenshots:

1. Sending a POST request to /orders via Postman

2. Logs of Docker Compose(Kafka, zookeeper and database) and Flask server

3. Exec into db and query the data added by consumer