Skip to main content

​
What are Alerts?

Alerts notify you when your agents exceed specified thresholds or meet certain conditions. Configure alerts to monitor costs, response times, and call volumes, ensuring you stay informed about critical events. Note: Email notifications require SMTP configuration during AgentOS installation. See the Installation guide for SMTP setup.

​
Alert Configuration

Create custom alerts to monitor your agents’ performance and resource usage.

​
Alert Parameters

When creating an alert, configure the following:
  • Alert Name - Descriptive name for the alert (min 3 characters)
  • Choose Agent - Select from deployed agents
  • Metric Name - What to monitor:
    • call - Number of API calls
    • cost - Total costs
    • response_time - Response duration
  • Condition - When to trigger:
    • lower_than - Alert when metric drops below threshold
    • equals_to - Alert when metric equals threshold
    • bigger_than - Alert when metric exceeds threshold
  • Time Interval - Evaluation period:
    • minute - Check per minute
    • hour - Check per hour
    • day - Check per day
  • Threshold Value - Numeric threshold for the condition
  • Email - Email address to receive notifications

​
Example Alert

“High Cost Alert”
  • Agent: My Production Agent
  • Metric: cost
  • Condition: bigger_than
  • Time Interval: day
  • Threshold: 100
  • Result: Alert when daily costs exceed $100
Alert Creation Form

​
How It Works

  1. Navigate to Alerts in the AgentOS dashboard
  2. Click Create New Alert
  3. Fill in the alert configuration form
  4. Save the alert
AgentOS will monitor your agent according to the specified parameters and send email notifications when conditions are met.

​
Alert Types

Common alert configurations:

​
Cost Monitoring

  • Track spending and get notified when costs exceed budget
  • Example: Alert when daily cost > $50

​
Performance Monitoring

  • Monitor response times and identify slowdowns
  • Example: Alert when response time > 5 seconds

​
Usage Monitoring

  • Track API call volumes and detect unusual activity
  • Example: Alert when hourly calls > 1000

​
Best Practices

  • Set realistic thresholds - Avoid alert fatigue from too many notifications
  • Use time intervals wisely - Daily for costs, hourly for performance
  • Monitor critical agents - Focus alerts on production agents
  • Test alerts - Verify email notifications work correctly