website · linkedin · github

Team Members

Project Goal

Develop an HTTP API for a team-member management application, with data persistence in Mongodb. All actions should be stored in the database.

Assumptions

Multi-user cases or authentication aren't implemented. Assuming all clients see the same list of team members.

Database Design

One user has multiple version based on update.

erDiagram
  User {
    ObjectId id
    string firstName
    string lastName
    string phone
    string email
    UserRole role
    Date createdAt
    Date updatedAt
  }

  UserVersion {
    ObjectId id
    string userId
    OperationType operation
    string firstName
    string lastName
    string phone
    string email
    UserRole role
    Date createdAt
    Date updatedAt
  }

  User ||--o{ UserVersion : "has"

Loading

Flow Diagram

To store versions, I have used change stream of MongoDB. All the operations are caught in the change stream and stored in another collection.

flowchart TB
    api(API)
    server(Server)
    mongo[(Mongo DB)]
    changestream(Change Stream)
    consumers(Consumers)

    api --> server;
    server --> mongo;
    mongo --Get row changes--> changestream;
    changestream --> consumers;
    consumers --Add new version--> mongo;

Loading

Installation

npm install --force

Running

This project requires a MongoDB cluster with replication enabled to activate change streams.

Docker

There is a docker-compose.yml file for starting Mongo cluster.

docker-compose up

Dropping MongoDB Collections

To drop collections in MongoDB, use the drop() method. This is required to avoid errors.

mongosh
use sample

db.user_versions.drop();
db.users.drop();

Run the project

Set the environment variables for the database name (DB_NAME) and URL (DB_URL):

export DB_NAME=sample
export DB_URL=mongodb://localhost:27017

Then, run Nest as usual:

npm start

Endpoint Details

Make sure to restart the service if you drop the collection.

Adding a Team Member

curl --location 'localhost:3000/users' \
--header 'Content-Type: application/json' \
--data-raw '{
    "firstName": "John",
    "lastName": "Doe",
    "email": "[email protected]",
    "phone": "+919999999999",
    "role": "admin"
}'

Editing a Team Member

curl --location --request PUT 'localhost:3000/users/:user_id' \
--header 'Content-Type: application/json' \
--data-raw '{
    "email": "[email protected]"
}'

Listing Team Members

curl --location 'localhost:3000/users'

Getting a Team Member

curl --location 'localhost:3000/users/:user_id'

Deleting a Team Member

curl --location --request DELETE 'localhost:3000/users/:user_id'

Getting a Team Member Versions

curl --location 'localhost:3000/users/:user_id/versions'

Future Improvements

We need to keep track of the pointer ID for changes in the change stream. Currently, it is missing. It can be synchronized in Redis. In the current scenario, server shutdowns can cause us to miss the changes if the data is modified.

To make it more resilient, we can implement Debezium + Kafka to directly sync changes to the database. This approach eliminates some implementations but comes with a higher learning curve and added expense. Additionally, it is less flexible.

Currently, all users have the ability to change anything, indicating a need for improvement. Additionally, implementing pagination could enhance speed. However, some indexes are missing, which is an area that could be improved.

Regarding quality assurance (QA), the addition of test cases is crucial. Although I've integrated the framework, the test cases are still missing. With this, proper error handling could be addressed.