Auto-Tuning Alert Thresholds
Auto-Tune analyzes your historical check data and incident history for a rule, then suggests or automatically applies threshold adjustments to reduce false positives. It's available on Professional and Team plans.
What Auto-Tune Does
Auto-Tune looks at:
- The historical distribution of the metric the rule monitors (response times, status codes, uptime)
- How often the rule has fired
- How often those firings resulted in meaningful incidents vs. false positives
From this data, it calculates threshold recommendations that would have reduced false positives while still catching real failures.
How to Use Auto-Tune
Getting a Suggestion
- Go to Alerts → Alert Rules.
- Open the rule you want to tune.
- Click Auto-Tune Threshold (or find it in the Alert Accuracy Dashboard under the rule's metrics).
- Review the suggested threshold.
- Click Apply Suggestion to accept it, or adjust manually and save.
From the Alert Accuracy Dashboard
- Go to Settings → Alert Accuracy.
- Find a rule with a high false positive rate.
- Click Tune next to the rule to see and apply the suggestion.
Reviewing Before Applying
Auto-Tune suggestions are based on historical data, not future predictions. Before applying:
- Review whether the suggested threshold makes sense for your service's SLA
- Consider whether recent infrastructure changes might have changed your baseline
- Apply with low-priority rules first to build confidence in the recommendations
Related Articles
Still have questions? Contact support.
Related Articles
Alert Accuracy Dashboard
The Alert Accuracy Dashboard tracks your team's false positive rate over time and helps you identify which alert rules are generating noise. It's available on Professional and Team plans. What It Shows The Alert Accuracy Dashboard gives you: ...
Testing an Alert Rule
The Test Rule feature lets you simulate your alert rule against current monitoring data to see what it would trigger — without creating a real incident or sending notifications. How to Test a Rule In the left sidebar, click Alerts, then Alert Rules. ...
Key Concepts: Endpoints, Checks, Incidents, and Alerts
This article explains the four core building blocks of PulseAPI and how they work together. Understanding these concepts makes every other part of the product easier to use. Endpoint (Monitor) An endpoint is a URL you want PulseAPI to watch. In ...
How Alert Rules Work
An alert rule defines the condition under which PulseAPI creates an incident and sends notifications. This article explains the evaluation flow, the components of a rule, and common configuration patterns. The Evaluation Flow After every check, ...
Alert Cooldowns and Deduplication
The cooldown period on an alert rule controls how frequently PulseAPI can fire that rule for the same monitor. This article explains how cooldowns prevent notification spam and how to set an appropriate value. The Problem Cooldowns Solve Without a ...