PulseGuard is a cloud-native microservices platform designed to monitor the liveness of background jobs, cron tasks, and backup scripts. Unlike traditional uptime monitors that ping a URL, PulseGuard waits for the monitored service to "check-in." If a signal is missed, an alert is triggered.

PulseGuard is built on a strict Microservices Architecture using Java 21 and the Spring Boot 3.5 / Spring Cloud 2025 ecosystem.

graph TD
    Client[Client / Script] -->|POST /ping| Gateway[API Gateway]

subgraph Infrastructure
Gateway -.->|Lookup| Eureka[Discovery Server]
Ingestion -.->|Register| Eureka
Processor -.->|Register| Eureka
end

subgraph Logic
Gateway -->|Route| Ingestion[Pulse-Ingestion Service]
Processor[Pulse-Processor Service] -->|OpenFeign / Sync| Ingestion
Ingestion -->|Writes| H2[(H2 Database)]
end

classDef service fill:#9f5f9f,stroke:#333,stroke-width:2px;
classDef infra fill:#6b996b,stroke:#333,stroke-width:2px;
classDef storage fill:#4d4d99,stroke:#333,stroke-width:2px;

class Ingestion,Processor service;
class Gateway,Eureka infra;
class H2 storage;

• Core: Java 21, Spring Boot 3.5.x
• Cloud Native: Spring Cloud Gateway, Netflix Eureka (Discovery)
• Communication: OpenFeign (Synchronous REST), RESTful APIs
• Data: Spring Data JPA, H2 (In-Memory / Pluggable)
• Tooling: Maven (Multi-Module), Docker, Docker Compose, Lombok

• Java 21 SDK
• Maven 3.8+
• Docker (Optional, for containerized deployment)

1. Clone the repository

  git clone https://github.com/stefanobini99/pulseguard.git
  cd pulseguard 

2. Build the project This project is a Maven Multi-Module build. Run the build from the root directory:

  mvn clean install

This spins up the entire infrastructure (Discovery, Gateway, Databases, Services) in an isolated network.

  docker-compose up --build -d

• Eureka Dashboard: http://localhost:8761
• API Gateway: http://localhost:8080

If you do not have Docker running locally, you can start the services individually in this specific order:

• Discovery Server: Run DiscoveryServerApplication
• API Gateway: Run ApiGatewayApplication
• Ingestion Service: Run PulseIngestionApplication
• Processor Service: Run PulseProcessorApplication

1. Send a Heartbeat (Ping)

   Scripts call this endpoint to confirm they are alive.

  curl -X POST http://localhost:8080/api/v1/ingestion \
     -H "Content-Type: application/json" \
     -d '{"monitorId": "backup-script-01", "status": "OK"}' 

2. View Dead Monitors (Internal/Debug)

   The Processor service calls this endpoint via OpenFeign to verify status.

    curl http://localhost:8080/api/v1/ingestion/dead-monitors

Currently, the getDeadMonitors() logic uses a Java Stream implementation for the MVP.

• Current State: Fetches dataset and filters in-memory (O(N)).
• Planned Upgrade: Migration to a custom JPQL/Native Query to handle the "Greatest-N-Per-Group" problem directly in the database engine for O(1) performance at scale.

• Implement Resilience4j Circuit Breakers on the Gateway.
• Migrate from H2 (In-Memory) to PostgreSQL for persistence.

Stefano Bini Java Software Developer | Spring Ecosystem Specialist