Network Gateway Case Study

Overview

The Network Gateway is a lightweight internal dashboard designed for a product team to monitor and access services across a Kubernetes-based development environment. It replaces scattered bookmarks and manual URL entry with a single, reliable interface for visibility and rapid access.

System Context & Problem

Operating on a single-master Kubernetes cluster, our services (Meilisearch, NATS, Kroki) were distributed across multiple ports. Before this tool, engineers relied on manual IP lookups, leading to repeated context switching and unnecessary interruptions between team members. The goal was not deep infrastructure control, but simple visibility and access.

Architecture & Performance

Built as a fully static, client-side application, the dashboard is deployable via Nginx with no backend dependency. This ensures the service directory remains accessible even if other internal systems fail.

Status and metrics are pulled using Prometheus running inside the cluster. By querying the Prometheus API directly via PromQL, the dashboard provides near real-time telemetry (CPU, Memory, Network) without the overhead of a dedicated backend server or heavy frontend frameworks.

Interface Design

The UI utilizes Progressive Disclosure. By default, the card-based layout allows for quick scanning of service health. Detailed metrics and pod-specific data are revealed only upon expanding a card, preventing information overload during time-sensitive situations.

A key feature is the Maintenance State (amber badges). This allows the team to communicate intentional downtime—such as "Down to save memory"—reducing false alarms and improving team coordination.

Outcome

The dashboard became the default entry point for the entire team, including the CEO and lead developer. By centralizing all endpoints and providing immediate state clarity, we have significantly reduced friction in our day-to-day development workflow.