Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.lumiqtrace.com/llms.txt

Use this file to discover all available pages before exploring further.

Alerts let you set thresholds on your agent metrics and receive a notification the moment one is crossed. Instead of checking the dashboard manually, you define the conditions that matter — a cost spike, an agent error rate surge, or a latency increase — and LumiqTrace notifies you automatically. Alert rules are evaluated every 5 minutes against your live data.
Alerts require the Pro plan or higher. The Free plan has no access to alert rules or alert history.

​
Plan limits

PlanAlert rulesWebhooks
FreeNone—
ProUp to 5 rulesNot available
TeamUnlimitedNot available
ScaleUnlimitedIncluded

​
Creating an alert rule

Click New rule to open the rule creation dialog.
1

Choose a metric

Select the metric you want to monitor:
  • error_rate — percentage of agent operations that fail, timeout, or are rate-limited
  • cost_usd — total agent spend in USD for the window
  • latency_ms — average response time in milliseconds
  • token_count — total tokens consumed
  • request_count — total operations
For agent-specific monitoring, use error_rate to catch tool failures and model errors, and cost_usd to guard against runaway agent loops.
2

Set the condition and threshold

Choose whether the rule fires when the metric is greater than (>) or less than (<) your threshold value.
3

Choose a time window

Select how long a window to aggregate before comparing to your threshold.
WindowBest for
5 minCatching sudden spikes from agent errors
15 minSustained error rate increases
30 minCost trend alerts
60 minLatency degradation patterns
4

Add notification recipients

Enter one or more email addresses to notify when the rule fires.
5

Add a webhook URL (Scale plan)

Enter a URL to receive a POST when the alert fires. See the Webhooks guide for the payload schema, signature verification, and integration examples with Slack and PagerDuty.
Click Save rule to activate it. It will be evaluated in the next 5-minute cycle.

​
Alert history

The alert history table shows the last 30 times any alert fired. Each entry shows the time, rule name, actual metric value, and configured threshold. If a rule fires very frequently, consider raising the threshold or increasing the time window to reduce noise. High-frequency alerts on error_rate often indicate an agent tool that fails intermittently — investigate the failing tool rather than silencing the alert.

​
Webhook payload

When a Scale plan alert fires and a webhook URL is configured, LumiqTrace POSTs a signed JSON payload: 
{
  "event": "alert.fired",
  "alert_id": "alert_abc123",
  "rule_name": "Agent error rate above 5%",
  "metric": "error_rate",
  "condition": ">",
  "threshold": 0.05,
  "current_value": 0.12,
  "window_min": 15,
  "project_id": "proj_xyz789",
  "organization_id": "org_def456",
  "fired_at": "2026-04-20T14:32:00Z",
  "dashboard_url": "https://app.lumiqtrace.com/your-org/your-project/traces"
}
Your endpoint should respond with a 2xx status code within 10 seconds. LumiqTrace retries failed deliveries up to 3 times with exponential backoff. See the full Webhooks guide for signature verification code and Slack/PagerDuty integration examples.
The dashboard_url field links directly to the traces page so whoever receives the alert can start investigating immediately without navigating to find the right project.