Probie

Realtime endpoint reachability monitoring for services, APIs, proxies, and network routes.

Probie helps you answer a simple question: when a tool keeps retrying, is the target service reachable from this machine, or is the problem more likely local network, proxy/VPN, DNS, or upstream instability?

中文说明: README.zh-CN.md

Install

npm install -g @rollpard/probie

For local development:

npm install
npm link

Quick Start

probie setup
probie watch

Enable system notifications for failures and recoveries:

probie watch --notify

Check once:

probie status

Profiles

Built-in profiles:

openai, anthropic, gemini, deepseek,
openrouter, groq, mistral, xai, perplexity, together, cohere

Monitor specific profiles:

probie watch --profile openai --profile anthropic

Save default profiles:

probie config set --profile openai --profile anthropic --interval 5
probie config show

Add a custom URL:

probie config add-url https://example.com/

Configuration

Tune timeouts and concurrency:

probie config set --connect-timeout 6 --max-time 10 --confirm-connect-timeout 20 --confirm-max-time 30 --concurrency 3

macOS autostart:

probie install-autostart
probie uninstall-autostart

How It Decides

• HTTP responses like 200, 401, 403, 404, and 405 count as reachable because the remote service responded.
• Network failures, DNS failures, TLS timeouts, and connection timeouts count as unreachable.
• VPN, system proxy, shell proxy, DNS, egress IP, country/region, and ASN are shown as context, not as the primary decision.
• --notify sends notifications when a selected service becomes unreachable and when it recovers.