Backup Cluster Manager (AMOS WS 2025/26)

Project Description

This project is an MVP developed for our industry partner SEP. It implements an alternative approach for managing distributed nodes in a backup cluster environment.

The system consists of:

• A Frontend that offers a UI for editing and viewing data from nodes, users, and their permissions.
• A Backend that simulates backup nodes and handles the interaction logic between nodes.

For comprehensive documentation, please visit our Project Wiki.

Components Overview

The project is divided into two main components. Please refer to their respective README files for detailed documentation and instructions on how to run them individually:

Frontend

The frontend is built with Angular. It communicates with the backend to get, edit, and delete data such as users, permissions, backups, and cluster information.

Backend

The backend is a Java Spring Boot application. It manages the business logic, simulates node interactions, and persists data.

Configuration

Database

The project uses a PostgreSQL database, which is configured via docker-compose.yml in the root directory.

Backend Configuration

Backend configuration can be found in backend/src/main/resources/application.yml.

Frontend Configuration

Frontend environment configurations are located in frontend/src/environments/.

Build and Run

Prerequisites

• Docker Desktop

Run Full Stack

To start the entire application (Frontend, Backend, and PostgreSQL), run the following command from the root directory:

docker compose up --build

This will start:

• Frontend at http://localhost:4200
• Backend (Cluster Manager) at http://localhost:8080
• PostgreSQL Database at http://localhost:5432

Run with Node Simulation

To start the application along with simulated nodes (defined in the test profile), use:

docker compose --profile test up --build

This includes the components above plus the configured backup nodes.

Metrics, Prometheus & Grafana

This project includes integrated monitoring with Prometheus and Grafana to track system performance metrics from the cluster manager and backup nodes.

Overview

• Prometheus: Scrapes and stores metrics from Spring Boot applications (Micrometer)
• Grafana: Visualizes metrics in dashboards with automatic provisioning
• Metrics: CPU, memory, heap usage, HTTP requests, garbage collection, and more

Running with Metrics

To start the stack with Prometheus and Grafana along with simulated nodes for more comprehensive metrics:

docker compose --profile test up --build

Accessing Prometheus

Visit http://localhost:9090 to access Prometheus:

• Targets: Shows all monitored instances
• Graph: Query and visualize metrics in real-time
• Status: View Prometheus configuration and alerts

Accessing Grafana

Visit http://localhost:3000 to access Grafana:

• Default login: admin / admin (change on first login)
• Data Source: Prometheus is pre-configured at http://prometheus:9090
• Dashboard: "Cluster Overview" dashboard is auto-loaded

Stress Testing

---

Requirement for running any stress test:

1. Navigate to the docker-compose.yml file located in the root directory of the project.
2. Replace the environment variable
   - SPRING_PROFILES_ACTIVE=cluster_manager
   inside the cluster-manager service with
   - SPRING_PROFILES_ACTIVE=cluster_manager,test-runner

Running the stress test:

1. Make sure you have followed the Requirement above for running any stress test.
2. Select a test from the /stress-test folder you want to run
   • Each test has its own folder with a descriptive folder name, starting with test-
   • Each test folder contains its own README.md file with specific instructions.
3. After selecting the test, run
   docker compose --profile [folder-name] up --build
   where [folder-name] is the name of the test folder you selected in step 2.
4. After running the test, you can find the results in the /stress-test/[folder-name]/results folder.
5. If you want to see the HTML report, run node generate-results-page.js which will update the results.html.

Credits

k6-reporter is used to create HTML reports of each test, you can access the code from Github