Wink

中文文档

A minimal, high-performance, single-binary self-hosted monitoring tool.

Lightweight alternative to Uptime Kuma — zero dependencies, file-based storage, single executable.

Features

• Single binary — Go backend + embedded frontend, no runtime dependencies
• File-based storage — JSON config and history, no database required
• HTTP / TCP / ICMP monitoring with configurable intervals
• Flapping control — debounce alerts with retry thresholds, no false alarms
• Reminder alerts — repeat notifications every N failures after DOWN
• Dynamic retry interval — faster probing when a monitor is failing
• Telegram & Webhook notifications with extensible notifier interface
• Notifier remark — label each notifier for easy identification in alert messages
• Inline notifier management — edit, test, and delete notifiers directly from settings
• Telegram Chat ID helper — fetch available chats from Bot API with one click
• Per-monitor notifier targeting — send alerts to specific notifiers only
• Monitor pause/resume — temporarily disable monitors without deleting them
• Grouped monitor list — monitors organized by group with collapsible sections
• Uptime tracking — 24h / 7d / 30d sliding window calculations
• Heartbeat bars — visual history of recent probe results per monitor
• Incident log — separate 30-day incident storage (incidents.json) with automatic cleanup
• Timezone support — auto-detects system timezone on first launch, configurable via UI
• SSO — reverse proxy Single Sign-On via Remote-User header
• Friendly error handling — inline toast notifications for form validation errors
• Atomic writes — crash-safe file persistence (write-sync-rename)
• Login rate limiting — per-IP lockout after failed attempts
• Session TTL — auto-expiring sessions with background cleanup
• Hot reload — add/edit/remove monitors without restart
• Web settings — configure system, auth, groups, and notifiers from the UI
• i18n — Chinese / English bilingual interface with one-click switching
• Dark mode — light / dark theme toggle
• Health endpoint — GET /healthz for external monitoring

Quick Start

Linux one-click install (recommended)

Requires: Linux (amd64/arm64), systemd, curl.

curl -fsSL https://raw.githubusercontent.com/makt28/wink/main/install.sh | sudo bash

After installation, use the wink command to manage the service:

sudo wink start       # Start service
sudo wink stop        # Stop service
sudo wink restart     # Restart service
sudo wink status      # Show status
sudo wink logs        # Follow logs (Ctrl+C to exit)
sudo wink update      # Update to latest release
sudo wink uninstall   # Remove Wink (preserves data)
sudo wink reinstall   # Re-download and restart

Data files are stored in /opt/wink/.

Download binary

Download the latest release for your platform from Releases:

Platform	File
Linux x86_64	wink-linux-amd64
Linux ARM64	wink-linux-arm64
macOS Apple Silicon	wink-darwin-arm64
Windows x86_64	wink-windows-amd64.exe

chmod +x wink-linux-amd64
./wink-linux-amd64

Termux (Android): Use wink-linux-arm64 — it runs natively on Termux's Linux kernel.

Build from source

Requires Go 1.24+ and Node.js (for Tailwind CSS compilation).

git clone https://github.com/makt28/wink.git
cd wink
npm install
make build
./wink

Cross-compile

Build for all supported platforms at once:

make cross

Outputs to dist/:

dist/wink-linux-amd64
dist/wink-linux-arm64
dist/wink-darwin-arm64
dist/wink-windows-amd64.exe

Login

Open http://localhost:8080 in your browser.

Default credentials: admin / 123456

Change the password in Settings > Auth after first login.

Configuration

See config.json.example for the full schema. Key sections:

Section	Description
system	Bind address, check interval, history limits, log level, timezone (auto-detected)
auth	Username, bcrypt password hash, login rate limiting, SSO toggle
contact_groups	Visual grouping for monitors
notifiers	Notification channels (Telegram, Webhook) with remark labels
monitors	List of targets to monitor (HTTP, TCP, Ping)

Monitor fields

Field	Description	Default
interval	Check interval in seconds	System default
timeout	Probe timeout in seconds	5
max_retries	Failures before marking DOWN	3
retry_interval	Faster interval when failing (0 = normal)	0
reminder_interval	Re-alert every N failures after DOWN (0 = off)	0
ignore_tls	Skip TLS certificate validation (HTTP only)	false
enabled	Enable/disable the monitor (null = true)	true
notifier_ids	Send alerts to specific notifiers only (empty = no notifications)	[]

Monitor types

Type	Target format	Example
http	Full URL	https://api.example.com/health
tcp	host:port	db.example.com:5432
ping	Hostname or IP	10.0.0.1

Note: Ping uses the system ping command — no special privileges needed. Make sure ping is available in your PATH.

Data files

File	Description
config.json	All configuration (system, auth, groups, monitors)
history.json	Latency history and uptime data per monitor
incidents.json	Incident records, auto-cleaned after 30 days

Development

# Run directly
make dev

# Format code
make fmt

# Static analysis
make vet

# Run tests
make test

# Full build (tailwind + fmt + vet + build)
make build

# Rebuild Tailwind CSS only
make tailwind

# Cross-compile all platforms
make cross

API

Health check

GET /healthz

Returns (no auth required):

{
  "status": "ok",
  "version": "0.1.5",
  "uptime_seconds": 86400,
  "monitor_count": 5
}

Architecture

Scheduler → 1 goroutine per monitor → Prober (HTTP/TCP/ICMP)
         → Analyzer (flapping control) → Notification Router → Telegram / Webhook
                                       → History Manager → history.json + incidents.json (atomic write)

• Frontend: Vanilla JavaScript + Tailwind CSS (compiled & embedded), all via go:embed
• Backend: Go + chi router, html/template rendering
• Storage: Atomic JSON file writes (write → sync → rename)

License

GPL-3.0