OKBlog

Welcome to OKBlog, where "OK" stands for "Over Kill"!

Why use a simple WordPress installation when you can create an army of microservices?

Is it necessary? Absolutely not.
Is it fun? Maybe?
Is it educational? I hope so!

When someone asks why you built a distributed system for a blog that gets twelve visitors a month, just say: "Because it's not alright, it's OKBlog!"

Architecture

The application consists of the following components:

• AI Service (Go/Gin): Provides AI Capabilities
• Profile Service (Go/go-kit): Manages user profiles with PostgreSQL database
• Post Service (Java/Spring Boot): Handles blog posts and comments with Kafka integration
• Search Service (Rust): Provides search functionality across the platform
• File Service (Python/Flask): Manages file uploads and storage using MinIO
• Tag Service (Go): Handles post categorization and tagging
• Web Service (Vue.js/Nuxt): Frontend application with SSR for better SEO
• Admin Service (React): Administrative interface for content management
• Nginx: Web server and API gateway
• Elasticsearch & Kibana: Centralized logging infrastructure

Getting Started

Prerequisites for development (optional)

• Go 1.21+ (for ai, profile and tag service development)
• Java 17+ (for post service development)
• Python 3.10+ (for file service development)
• Rust (for search service development)
• Node.js 18+ (for web and admin services development)

Running the Application

1. You need docker, docker compose, and python.

2. If you want, you can go to each folder and start the services, but I've prepared a running script to start all the services.

python ./tools/docker-container-manage/run-docker.py

3. Connect Debezium to MySQL and ElasticSearch.

python ./tools/init/init_debezium.py

4. Create Kibana Dataviews, which later can be accessed at http://localhost:5601

python ./tools/init/init_kibana.py

5. (Optional) Create a user and posts.

python ./tools/init/init_user_and_posts.py

Accessing Services

Public API

All requests are through the API Gateway, which can be accessed at http://localhost:80, while admin is accessed at http://localhost:3001

Internal API

• AI Service API: http://localhost:3003
• Profile Service API: http://localhost:8080
• Post Service API: http://localhost:8081
• Search Service API: http://localhost:8082
• File Service API: http://localhost:8083
• Tag Service API: http://localhost:8084
• Web Application: http://localhost:3000

Development

Each service can be developed independently:

• AI Service: See ai/README.md for details
• Profile Service: See profile/README.md for details
• Post Service: See post/README.md for details
• Search Service: See search/README.md for details
• File Service: See file/README.md for details
• Tag Service: See tag/README.md for details
• Web Service: See web/README.md for details
• Admin Service: See admin/README.md for details

CI/CD Workflows

The project uses GitHub Actions for continuous deployment of each microservice. The workflows are triggered when pushing tags with specific prefixes:

Service	Tag Format	Workflow File
AI Service	ai-*-*	ai-deploy.yml
Admin Service	admin-*-*	admin-deploy.yml
Profile Service	profile-*-*	profile-deploy.yml
File Service	file-*-*	file-deploy.yml
Post Service	post-*-*	post-deploy.yml
Search Service	search-*-*	search-deploy.yml
Web Service	web-*-*	web-deploy.yml
Tag Service	tag-*-*	tag-deploy.yml

How the Workflows Work

At the moment, deployment is using Caprover, and each workflow follows these steps:

1. Checks out the repository code
2. Build and run the tests
3. Sets up Docker Buildx for multi-platform builds
4. Logs in to DockerHub using repository secrets
5. Extracts the version from the tag name
6. Builds and pushes a Docker image to DockerHub
7. Deploys the service to CapRover using the appropriate app token

Triggering Deployments

To deploy a service, create and push a tag with the appropriate prefix:

# Example: Deploy admin service version 1-2
git tag admin-1-2
git push origin admin-1-2

Required Secrets

The workflows require the following secrets to be configured in your GitHub repository:

• DOCKERHUB_USERNAME - Your DockerHub username
• DOCKERHUB_TOKEN - Your DockerHub access token
• CAPROVER_SERVER - Your CapRover server URL
• CAPROVER_ADMIN_APP_TOKEN - CapRover token for admin service
• CAPROVER_AI_APP_TOKEN - CapRover token for ai service
• CAPROVER_PROFILE_APP_TOKEN - CapRover token for profile service
• CAPROVER_FILE_APP_TOKEN - CapRover token for file service
• CAPROVER_POST_APP_TOKEN - CapRover token for post service
• CAPROVER_SEARCH_APP_TOKEN - CapRover token for search service
• CAPROVER_WEB_APP_TOKEN - CapRover token for web service
• CAPROVER_TAG_APP_TOKEN - CapRover token for tag service