From host metricsto incident response.

A hybrid-cloud monitoring and observability platform connecting hosts, containers, metrics, logs, AWS resources, and incident workflows.

React / ViteNode.js / ExpressSocket.IOPostgreSQLAWS / EKSGitHub Actions

Explore the repository

Product: Monitoring and incident workflows
Delivery: GitHub Actions / OIDC / ECR
Deployment: Docker and Amazon EKS

SYSTEM OVERVIEW

Signals in. Operational context out.

Monitoring agents, the API, application data, and cloud integrations meet in one platform.

Application API

Node.js / Express / Socket.IO

Host agents
Metrics and host status
Console
React + Vite
Data
PostgreSQL / RDS
AWS
STS role integration

↳ Component overview · backend deployed and tested on EKS

The problem

Bring infrastructure health, incident state, and operational actions into a coherent interface. The deployment also needs to connect application networking, AWS identity, container delivery, and database configuration.

The build

Minerva pairs a React/Vite frontend with a Node.js/Express backend and Socket.IO. Host agents report operational metrics. Incidents support acknowledgement, resolution, and a timeline of detection, remediation, and comments. PostgreSQL provides persistence, with RDS-backed configuration for the AWS deployment.

Engineering decisions

Use explicit cloud identity

AWS accounts connect through STS AssumeRole with external IDs instead of storing long-lived AWS access keys. GitHub Actions uses OIDC for image delivery to ECR.

Follow connectivity into the cluster

A host agent initially could not reach the backend from its container. Correct service publishing and container-to-backend networking enabled metric delivery.

Debug the actual environment

A Kubernetes workload still targeted localhost for its database. Tracing that configuration moved the deployment toward an RDS-backed setup.

Keep incident state synchronized

Backend incident state, global incident counts, and browser events were connected so acknowledgement and resolution changes propagate through the interface.

WHAT THE WORK DEMONSTRATES

Built. Investigated. Applied.

Deployed backend

The backend was deployed and tested on EKS, extending the project beyond local Docker and EC2 environments.

Operational workflows

Host-agent metrics, container visibility, incident acknowledgement, resolution, and timelines connect monitoring to action.

Delivery and deployment lessons

GitHub Actions OIDC publishes images to ECR. Practical fixes addressed container networking, database configuration, and synchronized incident state.

Have an infrastructure problem worth solving?

Let’s build something useful