agentpulse-monitor（Agent 脉冲监视器）

AgentPulse — AI Infrastructure Monitoring

What It Does

AgentPulse monitors your servers with a thin, open-source agent and a smart cloud API. It learns your infrastructure's normal behavior over time, alerts you when something's wrong, and can auto-remediate issues with your approval.

Architecture: Dumb Agent, Smart Cloud

• The client agent is thin and open-source — ships metrics to the API, does zero analysis
• The cloud API receives metrics, runs analysis, stores history, sends alerts, handles remediation
• The moat is baseline learning (per-server norms over time) and accumulated intelligence

Quick Start

1. Register Your Server

curl -X POST https://api.agentpulse.io/v1/servers \
  -H "Content-Type: application/json" \
  -d '{"hostname": "my-server", "plan": "starter"}'

Save the returned server_id and api_key.

2. Install the Agent

One-liner install:

curl -fsSL https://agentpulse.io/install.sh | bash -s -- --api-key YOUR_API_KEY --server-id YOUR_SERVER_ID

Or manual install:

# Download
wget -O /usr/local/bin/agentpulse-agent https://agentpulse.io/downloads/agent_client.py
chmod +x /usr/local/bin/agentpulse-agent

# Configure
cat > /etc/agentpulse.conf << 'EOF'
API_KEY=your_api_key_here
SERVER_ID=your_server_id_here
API_URL=https://api.agentpulse.io
EOF

# Run a quick health check
agentpulse-agent --quick

3. Set Up Cron

# Quick check every 5 minutes
*/5 * * * * /usr/local/bin/agentpulse-agent --quick

# Full report every 30 minutes
*/30 * * * * /usr/local/bin/agentpulse-agent --full

# Daily deep report at 8am UTC
0 8 * * * /usr/local/bin/agentpulse-agent --full

4. Get Alerts

Alerts are delivered via Telegram, email, or webhook. Configure your notification channels in the dashboard at https://agentpulse.io/dashboard.

What Gets Monitored

Critical Services (immediate alert if down)

• SSH, Docker, Nginx, your custom services

System Metrics

• CPU, memory, disk usage with trend analysis
• Process counts and zombie detection
• Load average and uptime tracking

Network

• Listening ports (unexpected changes flagged)
• Active connections
• DNS resolution health

Baseline Learning (after 24h)

• Per-server normal ranges for all metrics
• Anomaly detection when values deviate from learned norms
• Trend alerts before thresholds are breached

Alert Levels

Auto-Remediation (Pro + Business)

When a critical alert fires, AgentPulse can suggest or execute remediation actions:

• Restart crashed services
• Clear disk space (logs, caches)
• Rotate logs automatically
• Scale resources

All auto-remediation requires your approval via Telegram buttons or dashboard confirmation.

Pricing

API Reference

Submit a Report

POST /api/v1/report
Headers: X-API-Key: your_api_key
Body: { "server_id": "...", "metrics": {...} }

Get Server Status

GET /api/v1/servers/{server_id}/status
Headers: X-API-Key: your_api_key

Get Alert History

GET /api/v1/servers/{server_id}/alerts?limit=50
Headers: X-API-Key: your_api_key

Health Check

GET /health

Requirements

• Linux server (Ubuntu 20.04+, Debian 11+, CentOS 8+)
• Python 3.8+
• curl or wget
• Cron (for scheduled reports)

Support

• Docs: https://agentpulse.io/docs
• Telegram: @agentpulse_support
• Email: support@agentpulse.io