Procboss (pboss) — a blazing-fast, universal production process manager built on Bun native APIs. Run, cluster, monitor, and manage any application — Node.js, Bun, Go, Python, Rust, Ruby, PHP, Java, native binaries, and shell scripts — with pure performance and zero overhead. Open source (GPLv3), by procboss.com.
ProcBoss Cloud (procboss.com) — a hosted layer on top of pboss: link the servers you already run, and get fleet visibility, alerts, and metrics from one dashboard.
Why pboss?
pboss replaces complex, heavyweight process managers with a clean, ultra-fast architecture. It uses Bun.spawn for lightning-fast process orchestration, Bun.serve for the real-time web dashboard and IPC, native WebSocket over Unix sockets, Bun.file for high-performance I/O, and Bun.gzipSync for automatic log compression. The result is a single machine-level daemon that starts in under 50ms, uses only ~12MB of RAM, and manages your entire infrastructure seamlessly.
Highlights
Universal runtime support — auto-detection for Node.js, Bun, Go, Python, Rust, Ruby, PHP, Java JARs, shell scripts, and compiled binaries.
Cluster mode — multiple instances with per-worker env injection, automatic port assignment, and zero-downtime rolling reloads.
Namespaces — group related processes (--namespace my-app) and operate on the whole group: pboss restart my-app, pboss stop my-app, pboss start my-app, pboss delete my-app (confirmed; --force skips). Unknown targets are clear errors, and group operations say what they touched.
Foreground mode — --no-daemon blocks as PID 1, purpose-built for Docker, Kubernetes, and containers.
Web dashboard — self-contained, live WebSocket updates, CPU/memory charts, process controls, and a log viewer. No external dependencies.
Prometheus metrics — dedicated /metrics endpoint, ready for scraping and Grafana dashboards.
Log management — automatic capture, size-based rotation, retention, and optional gzip compression.
Health checks, cron restarts, file watching — keep processes healthy and self-healing.
Standalone cron jobs — schedule any command with friendly syntax (pboss cron run everyday@9:11 "bun backup.ts", every-second to on-date@24-10-2026-23:10), no managed process required; persists across reboots.
Persistence (default on) — the process list is saved automatically after every change, and the boot service (systemd / launchd / Task Scheduler) is installed automatically at install time — your apps survive daemon restarts and system reboots out of the box.
Remote deployment — SSH-based deploys with release directories, symlink rotation, and pre/post hooks.
Tiny footprint — a single machine-level daemon that starts in under 50ms and uses only ~12MB of RAM.
