Integration Guide
Get up and running in under 2 minutes.
1Quick Start
python
# Install
pip install agent-monitor
# Initialize
from agent_monitor import Monitor
monitor = Monitor(api_key="aw_live_your_key_here")
# Track events
monitor.track("my-agent", event="task_complete", metadata={
"tokens": 150,
"latency_ms": 340,
"model": "gpt-4"
})
# Heartbeat (call periodically)
monitor.heartbeat("my-agent")
# Error tracking
try:
result = agent.run(task)
monitor.track("my-agent", event="success", metadata={"result": result})
except Exception as e:
monitor.track("my-agent", event="error", metadata={"error": str(e)})2Webhook Configuration
Receive real-time event notifications via webhooks. Configure in your dashboard or via API:
POST /api/v1/webhooks
{
"url": "https://your-server.com/webhooks/agentwatch",
"events": ["agent.down", "agent.error", "agent.budget_exceeded"],
"secret": "whsec_your_signing_secret",
"headers": {
"X-Custom-Header": "value"
}
}Available Events
agent.upagent.downagent.erroragent.recoveredagent.budget_exceededagent.latency_spike
3Alert Rules
Define custom alert rules using simple YAML syntax or the dashboard UI:
agentwatch.yaml
rules:
- name: "Agent Down"
condition: heartbeat_missing > 60s
severity: critical
notify: [slack, email]
- name: "High Error Rate"
condition: error_rate > 5%
window: 5m
severity: warning
notify: [slack]
- name: "Budget Alert"
condition: daily_tokens > 500000
severity: warning
notify: [email]4API Reference
POST/api/v1/trackSend an event for an agent
POST/api/v1/heartbeatSend agent heartbeat
GET/api/v1/agentsList all agents
GET/api/v1/agents/:id/metricsGet agent metrics
GET/api/v1/alertsList recent alerts
POST/api/v1/webhooksCreate webhook endpoint