What is this?

This site showcases 5 interactive tech demos simulating backend challenges for enterprise services at the scale of 10K to 1M users. Each demo runs a real-time data simulator visualizing architecture concepts that are usually only discussed on whiteboards.

How to use

Click any card below to enter a demo page, then press "Start Demo" to see real-time dynamic visualizations.

Tech Stack

Backend Go 1.22+ (goroutine / WebSocket / simulated data), Frontend Next.js + shadcn/ui + Canvas charts.

Scenario

10K-1M users, 5-person team, no overtime allowed, fully self-hosted, ISO 27001:2022 compliant.

Enterprise Architecture Showcase

Go + Reactfor 100K Users

Interactive demonstrations of enterprise-grade backend patterns. Each demo runs a real Go backend with WebSocket streaming to a React frontend, visualizing concepts that are typically only discussed in whiteboards.

Go 1.22+ Backend

Next.js Frontend

WebSocket Real-time

K8s Ready

High Concurrency

Goroutine Pool + Rate Limit + Singleflight

Watch goroutines spawn, worker pools process tasks, rate limiters reject excess traffic, and singleflight merge duplicate requests in real time.

goroutinesemaphoresingleflightworker pool

Distributed Monitoring

RED Metrics + Alerting + Trace Chain

Live dashboard showing request rate, error rate, latency percentiles across 5 microservices. Simulates anomaly injection with alert escalation and distributed trace visualization.

PrometheusOpenTelemetrySLOerror budget

Real-time Collaboration

CRDT + WebSocket + Conflict Resolution

Two editor panes synced via Yjs CRDT over WebSocket. Type in one, see it appear in the other. Demonstrates the core technology behind Notion-like collaborative editing.

YjsCRDTWebSocketTipTap

Database Pool Management

Connection Pool + Slow Query + Saturation

Visualize PostgreSQL connection pool state: active/idle connections, queue depth, saturation alerts. Includes live slow query ranking from pg_stat_statements.

PostgreSQLconnection poolpg_stat_statementsPITR

Kubernetes Operations

HPA + Rolling Update + OOMKilled + Bin-packing

Watch pods scale up/down with HPA, rolling updates replace versions one by one, OOMKilled recovery, and node resource bin-packing across a 3-node cluster.

HPARolling UpdatePDBOOMKilledbin-packing

Architecture Context

Scale

100K internal users, ~60K DAU, peak 30K QPS. Internal network, no public internet exposure.

Team

5-person team (1 lead + 2 backend + 2 frontend). Modular monolith first, microservices when needed.

Compliance

ISO 27001:2022 ready. All infrastructure self-hosted. GitOps + SAST/SCA in CI pipeline. Zero data egress.