Threshold Alerts
Threshold alerts let you define the performance standards your API must meet. When a test result crosses a threshold, Load Curl sends an alert — so you catch regressions before they reach production.
:::info Plan availability Threshold alerts are available on the Pro plan and above. :::
Creating a threshold
- Open Settings → Alerts in the sidebar.
- Click New Threshold.
- Fill in the threshold form:
| Field | Description | Example |
|---|---|---|
| Metric | The metric to watch | P95 Latency |
| Operator | Condition direction | greater than |
| Value | The limit to enforce | 400 ms |
| Scope | Apply to all tests or a specific project | Production API |
| Notify | Email addresses to alert | team@yourapp.com |
- Click Save Threshold.
The threshold is now active. Every time a test in the selected scope completes, Load Curl evaluates all thresholds against the results.
Available metrics for alerting
You can create thresholds on any of the following:
- P50 / P95 / P99 latency (milliseconds)
- Error rate (percentage)
- Availability (percentage)
- Requests per second (floor — alert when RPS drops below a value)
- Server CPU peak (percentage)
- Overall report card grade (e.g., alert if grade drops below B)
Notification channels
Email (available now)
Every threshold alert sends an email to the addresses you configure. The email includes:
- Which threshold was breached
- The actual value vs. the threshold value
- A direct link to the Report Card
- The grade change from the previous run (if any)
PagerDuty and additional channels (roadmap)
PagerDuty integration and further alert channels (Slack, webhooks) are planned for an upcoming release.
Using thresholds in CI/CD
Thresholds integrate directly with the CLI and CI/CD pipeline. When a threshold is breached, the lf test command exits with a non-zero status code, which causes the pipeline step to fail.
# GitHub Actions example
- name: Load test
run: |
lf test \
--endpoint https://api.yourapp.com/v2/users \
--workers 500 \
--duration 60s \
--fail-on-threshold
The --fail-on-threshold flag checks all thresholds you have configured for the project. If any are breached, the test is marked as failed and the deployment is blocked.
See CI/CD Integration for the full pipeline setup guide.
Managing existing thresholds
From Settings → Alerts you can:
- Edit any threshold by clicking the row
- Pause a threshold temporarily (e.g., during planned maintenance)
- Delete a threshold you no longer need
- View alert history — a log of every time a threshold fired, with timestamps and the metric value that triggered it
Next step
Learn how to test a full user journey with the Scenario Builder.