BM2 - Bun Process Manager

A blazing-fast, full-featured process manager built entirely on Bun native APIs. The modern PM2 replacement — zero Node.js dependencies, pure Bun performance.

PM2 is the de facto process manager for Node.js, but it carries years of legacy baggage, a heavy dependency tree, and is fundamentally built for the Node.js runtime. BM2 is a ground-up reimagining of production process management designed exclusively for the Bun runtime.

BM2 replaces PM2's Node.js internals with Bun-native APIs. It uses Bun.spawn for process management, Bun.serve for the dashboard and IPC, native WebSocket for daemon communication, Bun.file for high-performance I/O, and Bun.gzipSync for log compression. The result is a process manager that starts faster, uses less memory, and leverages Bun's superior performance across the board.

Features Core Process Management — Start, stop, restart, reload, delete, and scale processes with automatic restart on crash, configurable restart strategies, memory-limit restarts, and tree killing.

Cluster Mode — Run multiple instances of your application with per-worker environment injection, automatic port assignment, and round-robin-ready configuration using NODE_APP_INSTANCE conventions.

Zero-Downtime Reload — Graceful reload cycles through instances sequentially, starting the new process before stopping the old one, ensuring your application never drops a request.

Foreground / No-Daemon Mode — Run BM2 in blocking foreground mode without spawning a background daemon. Designed for containerized environments like Docker, Kubernetes, and any platform that expects PID 1 to remain in the foreground.

Real-Time Web Dashboard — A built-in dark-themed web dashboard with live WebSocket updates, CPU/memory charts, process control buttons, and a log viewer. No external dependencies.

Prometheus Metrics — A dedicated metrics endpoint exports process and system telemetry in Prometheus exposition format, ready for scraping by Prometheus and visualization in Grafana.

Log Management — Automatic log capture with buffered writes, size-based rotation, configurable retention, optional gzip compression, log flushing, and real-time tailing.

Health Checks — HTTP health check probes with configurable intervals, timeouts, and failure thresholds that automatically restart unhealthy processes.

Cron Restarts — Schedule periodic restarts using standard cron expressions for applications that benefit from regular recycling.

File Watching — Automatic restart on file changes with configurable watch paths and ignore patterns. Ideal for development workflows.

Ecosystem Files — Declare your entire application topology in a single JSON or TypeScript configuration file and start everything with one command.