Docs » µAPM Detectors and Alerts » Building Rules for µAPM Detectors

Building Rules for µAPM Detectors

Note

This section explains how to build detector rules when you are creating a µAPM detector . If you are creating an Infrastructure or Custom Metrics detector, see this section instead.

Overview

This page explains how to build rules for detectors to monitor µAPM metrics. It assumes you are generally familiar with detectors and have already decided how you want to create the detector.

  • In the Alert signal tab, you select whether you want the detector to monitor latency or errors.
  • In the Alert condition and Alert settings tabs, you specify criteria for triggering an alert.
  • In the Alert message and Alert recipients tabs, you specify who should receive notifications and add notes that will be included in the notifications. In the Activate tab, you name the rule and make the detector “live.”

After you activate the detector, it will begin monitoring the signal immediately. When the signal meets the specified criteria, the detector will trigger alerts, create events, and send the specified message to the alert recipients.

Each tab is discussed below.

Alert signal

In the Alert signal tab, you choose whether you want to monitor latency or errors, and then specify the service endpoint(s) to monitor. If you are using multiple clusters in your Smart Gateway, you can also choose the cluster on which you want to set a detector.

If you have created the detector from the µAPM Services page, the relevant service, endpoints, and cluster will be filled in automatically. You can edit this information and add additional filters while you are creating the detector.

Endpoints are monitored using OR logic. In other words, the detector triggers an alert if any selected service/endpoint/cluster combination meets the trigger condition you specify.

Alert condition

In the Alert condition tab, you select the type of condition that will trigger an alert.

SignalFx provides several built-in alert conditions to make it simple for you to create robust alert conditions without needing to build advanced conditions behind the scenes.

The following table summarizes the available built-in µAPM alert conditions for Latency detectors.

Name Description Summary (samples)
Static threshold Alerts when latency goes above a static threshold, relative to a specified percentile, for a specified period of time. The 90th percentile of latency is above 500ms for 100% of 5 seconds.
Sudden change 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.

Latency in the last 10 minutes is more than 5 deviations above the norm established in the preceding 1 hour.

The 50th percentile of latency in the last 10 minutes is more than 30% above latency of the preceding 1 hour.

Historical anomaly Alerts when latency anomalously spikes compared to the same periods in the past (for cyclical or seasonal data). Anomaly can be defined through number of deviations from historical norm or percentage change compared to historical norm.

Latency in the last 10 minutes is more than 5 deviations above its historical norm, cyclical over 1‑week periods.

The 90th percentile of latency in the last 10 minutes is more than 30% above its historical norm, cyclical over 1‑week periods.

The following table summarizes the available built-in µAPM alert conditions for Error Rate detectors.

Name Description Summary (samples)
Static threshold Alerts when the error rate goes above a specified percentage for a minimum number of requests. The error rate over the last 10 minutes is above 10% across at least 50 requests.
Sudden change Alerts when there is a sudden increase in error rate. The error rate over the last 5 minutes is over 5% larger than the error rate of the preceding 1 hour across at least 100 requests.

After you have selected the alert condition, continue to the Alert Settings tab to specify the settings that will trigger alerts.

Alert settings

In the Alert settings tab, you specify the settings that will trigger an alert. The available settings vary depending on the alert condition you selected. For information on settings available for each condition, use the links in the table above.

Alert message, alert recipients, and activating the detector

After you have specified settings for triggering alerts, you can specify an alert message and alert recipients, and then activate the detector. These steps are the same for both APM and Infrastructure detectors. For more information on each tab, see Alert message, Alert recipients, and Activate.