kewinremy/rapi

RAPI

Context

Rapi is a simple application illustrating how to create an api with golang, gin & postgres.

Running the application

1. start the postgres database: sudo service postgresql start
2. build the application: go build
3. run the application: go run rapi

Endpoints & Usage

GET /v1/ping

This endpoint is used to check if the application is running.

Example

curl http://localhost:8888/ping

{
  "message": "pong"
}

GET /v1/items

This endpoint is used to get all items.

Example

curl  -X GET \
      http://localhost:8888/v1/items

response

[
    {
        "id": 1,
        "name": "copegus",
        "reservation_id": "986"
    },
    {
        "id": 2,
        "name": "esbriet",
        "reservation_id": "652"
    },
    {
        "id": 3,
        "name": "fuzeon",
        "reservation_id": "453"
    }
]

POST /v1/items

Because the reservation service has a delay (I added here 10 seconds for debugging), the http response is sent back before the item entity is created in the database. In the case of a web application, there shall be a way to communicate this change to the user.

Example

curl  -X POST \
	-H "Content-Type: application/json" \
      -d '{"name": "item_name"}' \
      http://localhost:8888/v1/items

logs

[GIN] 2022/09/28 - 07:01:56 | 200 | 112.71µs | 127.0.0.1 | POST "/v1/items"
2022/09/28 07:01:56 Sleeping for 10 seconds... 
2022/09/28 07:02:06 reservationIdChannel:  81
2022/09/28 07:02:06 1 rows inserted

Development

According to go idiomatic guidelines:

• https://go.dev/doc/faq7
• https://go.dev/ref/spec

Missing parts in application code

• config management for environment variables
• dependency injection library
• lazy object instantiation
• generic service library for CRUDO operations
• refactoring using the gorm orm: https://gorm.io/docs/
• enhancing the middleware with security and logging features (jwt, etc.)
• introducing editorconfig for code formatting

Missing parts in the infrastructure

• setting up the dev container with all required components (db + app)
• writing helm charts of the application
• orchestrating the deployment of the microservice with kubernetes (db & api)
• ci / cd pipeline with concurse ci (https://concourse-ci.org/pipelines.html)
• introduction of sonarqube for code scanning
• introduction of monitoring tools (prometheus, grafana, etc.)

Parts that are not there but not missing

• Pure OOP
• Adding new packages for components which are not reusable
• Introduction of data mappers between dtos & models only at a later stage of the project when repositores & services are diverging.