MongoDB Query "Top"

A modern, full-stack MongoDB monitoring tool that presents db.currentOp() results in an intuitive interface. Built with TypeScript in a monorepo architecture, featuring a REST API with Server-Sent Events and a React web dashboard.

• Real-time monitoring with auto-refresh and SSE streaming
• Intelligent filtering of system/internal queries
• Color-coded highlighting for unindexed queries (COLLSCAN)
• GeoIP location display for public IPs
• Auto-save long-running and problematic queries
• Interactive controls (pause, reverse, snapshot, show all)
• Multi-server support with connection management
• Web dashboard with virtualized table and JSON viewer

The built-in db.currentOp() has limitations:

• ❌ JSON output not easily readable
• ❌ Cluttered with system queries
• ❌ No auto-refresh or persistence
• ❌ No summary statistics

This tool provides:

• ✅ Human-readable tabular display
• ✅ Automatic filtering of noise
• ✅ Auto-refresh and real-time streaming
• ✅ Instant identification of slow queries
• ✅ Detection of unindexed scans
• ✅ REST API and web dashboard

pnpm install

# Configure your MongoDB servers (see Configuration section)
cp config/local.yaml.example config/local.yaml

# Start web dashboard
pnpm run dev:web

Run both API and Web services in containers:

# 1. Configure MongoDB and API settings in config/local.yaml
cp config/local.yaml.example config/local.yaml

# 2. Build Docker images (generates config from YAML - no env vars!)
pnpm docker:build

# 3. Start services
docker compose up -d

Access: Web UI at http://localhost:9000, API at http://localhost:9001

Production: pnpm docker:build https://api.yourdomain.com

See docs/DOCKER.md for complete documentation including production deployment, SSE-capable reverse proxy configuration, troubleshooting, and more.

pnpm run dev:web

Opens http://localhost:9000 with:

• Real-time query monitoring
• Interactive table with virtualization
• Query details with JSON viewer
• Server selection and connection management

# Start API only on http://localhost:9001
pnpm run dev:api

See docs/API.md for detailed API documentation.

Uses YAML configuration files in config/:

config/default.yaml (checked into git):

servers:
    localhost:
        name: Local MongoDB
        uri: mongodb://localhost:27017

api:
    port: 9001
    host: 0.0.0.0
    logLevel: info
    apiKey: dev-key-change-in-production
    cors:
        origins:
            - http://localhost:9000
        credentials: true

config/local.yaml (gitignored - your servers):

servers:
    production:
        name: Production Cluster
        uri: mongodb+srv://user:pass@cluster.mongodb.net/db

    staging:
        name: Staging
        uri: mongodb://user:pass@staging:27017/db?authSource=admin

# Override API settings
api:
    apiKey: your-secure-api-key-here
    logLevel: debug

Copy config/local.yaml.example to get started.

Monorepo Structure (Turborepo + pnpm workspaces):

apps/
├── api/      # Fastify REST API + SSE streaming
│             # Includes MongoDB services and query processing
└── web/      # React dashboard (TanStack Router, Zustand, shadcn/ui)
│             # Includes utility functions for styling

packages/
└── types/    # Shared TypeScript types

Services (apps/api/src/core):

• MongoConnectionService - Connection pooling
• QueryService - Query processing and filtering
• QueryLoggerService - Logging and snapshots

• Monorepo: Turborepo, pnpm workspaces
• Backend: TypeScript, MongoDB Driver v7, Fastify, lodash-es
• Frontend: React 19, TanStack Router + Virtual, Zustand, Vite, Tailwind CSS, shadcn/ui

# Install dependencies
pnpm install

# Development modes
pnpm run dev:api    # API only
pnpm run dev:web    # API + Web (recommended)
pnpm run dev        # All apps

# Build all packages
pnpm run build

# Build specific package
turbo build --filter=@mongo-query-top/api

# Format code
pnpm run format

# Production
pnpm run start:api

• docs/API.md - Complete API endpoint documentation
• docs/DOCKER.md - Docker deployment guide with production setup and SSE support
• CLAUDE.md - Developer guide with code patterns, architecture details, and customization instructions

Queries are auto-saved to logs/<server-id>/ when:

• Runtime exceeds configured threshold (default: 10s)
• Query uses COLLSCAN (collection scan)

MIT

Rushi Vishavadia