Monitoring Lookout for Metrics with Amazon CloudWatch

When an anomaly detector processes data or sends an alert, Amazon Lookout for Metrics sends metrics to Amazon CloudWatch. You can build graphs and dashboards with these metrics in the CloudWatch console to monitor detector activity, and set alarms that notify you when there are errors in alert delivery or anomaly detection. To filter and sort metrics, use the AlertArn and AnomalyDetectorArn Amazon CloudWatch dimensions.

To view metrics in the CloudWatch console

1. Open the Amazon CloudWatch console Metrics page (AWS/LookoutMetrics namespace).

2. Choose a dimension schema.

   • AlertArn – To view metrics for an alert channel.

   • AnomalyDetectorArn – To view processing metrics for a detector.

3. Choose metrics to add them to the graph.

4. To choose a different statistic and customize the graph, use the options on the Graphed metrics tab. By default, graphs use the Average statistic for all metrics.

Pricing

CloudWatch has an Always Free tier. Beyond the free tier threshold, CloudWatch charges for metrics, dashboards, alarms, logs, and insights. For details, see CloudWatch pricing.

For more information about CloudWatch, see the Amazon CloudWatch User Guide.

Using processing metrics

When a detector processes data at the end of each interval, it sends metrics to Amazon CloudWatch. You can use these metrics to monitor detector activity and trigger an alarm if a detector fails to process data.

There are three processing metrics:

• ExecutionsStarted – The number of times the detector looked for anomalies. For an active detector, this metric is sent once per interval.

• ExecutionsSucceeded – The number of processing attempts that succeeded.

• ExecutionsFailed – The number of processing attempts that failed.

View the processing metrics with the Sum statistic.

Using alert metrics

When a detector finds an anomaly, it sends anomaly alerts to alert channels that you configure. Each time it sends an anomaly alert, the detector sends a metric to indicate whether the alarm was sent successfully.

There are two alert metrics:

• Delivered – An anomaly alert was sent successfully.

• Undelivered – An anomaly alert could not be delivered.

View the alert metrics with the Sum statistic.

Configuring alarms

To get notifications when a metric exceeds a threshold, create an alarm. For example, you can create an alarm that sends a notification when the sum of the ExecutionsFailed metric exceeds 1 for an hour.

To create an alarm

1. Open the Amazon CloudWatch console Alarms page.

2. Choose Create alarm.

3. Choose Select metric and locate a metric for your detector, such as ExecutionsFailed for my-detector.

4. Follow the instructions to configure a condition, action, and name for the alarm.

For detailed instructions, see Create a CloudWatch alarm in the Amazon CloudWatch User Guide.