Docs » Configure detectors and alerts in SignalFx µAPM » Use default alert conditions in SignalFx µAPM » Latency - sudden change

Latency - sudden change 🔗

Note

This page explains how to use this condition when you are creating a µAPM detector. If you are creating an Infrastructure or Custom Metrics detector, see this page instead.

What this alert condition does 🔗

Alerts when latency during a recent time window anomalously spikes compared to the preceding time window. Anomaly can be defined through number of deviations from norm or percentage change of signal.

When to use this alert condition 🔗

Use this condition to find a sudden change in latency when you don’t have a static baseline for comparison. For example, a code push might cause a sudden increase in latency.

Basic settings 🔗

PARAMETER VALUES USAGE NOTES
Trigger Sensitivity Low, Medium, High, Custom Approximately how often alerts will be triggered, where Low can result in fewer alerts being triggered and alerts taking longer to clear (least flappy). Choose Custom to modify the settings that determine triggering and clearing sensitivity (listed below).

Advanced settings 🔗

PARAMETER VALUES USAGE NOTES
Alert based on Deviations from norm, Percentage change How to define anomalies as compared to the historical window.
Percentile to monitor (when Alert based on is Percentage change) 50, 90, 99 The percentile to monitor. For example, a value of 90 measures changes in the 90th percentile latency value.
Current window Integer >= 1, followed by time indicator (s, m, h, d, w), e.g. 30s, 10m, 2h, 5d, 1w The time range being monitored. For example, a value of 10m indicates that you are looking for anomalies that have occurred in the most recent 10 minutes. Should be significantly higher than signal resolution.
Historical window Integer >= 1, followed by time indicator (s, m, h, d, w), e.g. 30s, 10m, 2h, 5d, 1w. Must be higher than Current window. The time range being used to define the recent trend. For example, a value of 8h indicates that you are comparing values received during the current window with values received during the preceding 8 hours. Should be significantly higher than signal resolution.
Trigger threshold and Clear threshold (when Alert based on is Deviations from norm) Number >= 0.

Trigger the alert when the current value is greater than this number of deviations above historical norm. Higher trigger values result in lower sensitivity and potentially fewer alerts.

Clear the alert when the current value is less than this number of deviations above historical norm. Higher clear values result in alerts clearing more quickly.

Trigger threshold and Clear threshold (when Alert based on is Percentage change) Number between 0 and 100, inclusive; Clear threshold must be lower than Trigger threshold.

Trigger the alert when the current value is more than this percentage above the historical norm. For example, a trigger value of 30 will trigger an alert when the values being compared differ by more than 30%. Higher values result in lower sensitivity and potentially fewer alerts.

Clear the alert when the current value is less than this percentage above the historical norm. For example, a clear value of 20 will clear the alert when the values being compared differ by less than 20%. Higher values result in alerts taking longer to clear.

Exclude errors? Yes, No Whether to include or exclude error spans from the calculations.
Min. req/sec (absolute) Number Minimum requests/second required in current window to trigger or clear the alert; prevents alerts for sparse data.
Min. req/sec (% of history) Number between 0 and 100, inclusive Minimum requests/second, as a percentage of the historical rate, required in current window to trigger or clear the alert; prevents alerts for sparse data.