Amazon Lookout for Metrics FAQs

Amazon Lookout for Metrics uses machine learning (ML) to make it easier for customers to accurately detect anomalies in their metrics. After a customer uploads their data, Lookout for Metrics will automatically inspect it and uses ML to create accurate anomaly detection models. When anomalies are detected, Lookout for Metrics will group related anomalies together and provide a severity score so customers can diagnose issues quickly or maximize opportunities.

Amazon Lookout for Metrics has four key differentiated capabilities.

• Automates anomaly detection on data using ML: Lookout for Metrics supports a collection of machine learning algorithms. No one algorithm works for all kinds of data, so Lookout for Metrics inspects the data and applies the right ML algorithm to the right data to detect anomalies accurately. Because Lookout for Metrics takes care of the heavy lifting associated with selecting, training, building, and deploying the models, customers do not need any machine learning experience to get started quickly.
• Helps customers get started with little to no data: Customers can provide historical data for training the models if they choose. In the event customers choose not to or don’t have historical data, Lookout for Metrics can learn as it goes. The amount of time for Lookout for Metrics to learn and detect anomalies will vary based on the customer’s data.
• Provides actionable results: Lookout for Metrics will group concurrent anomalies into logical groups and send a single alert for the anomaly group rather than separate alerts, enabling customers to see the full-picture. The anomalies are also ranked in order of severity so that customers know which ones to focus on immediately.
• Continuously improves via human-in-the-loop feedback: Lookout for Metrics allows customers to provide feedback on the detected anomalies, which is used to improve the accuracy and performance of the models continuously.

Lookout for Metrics supports anomaly detection use cases across a wide variety of business metrics related to customer engagement, operations, sales, and marketing in industries like retail, gaming, ad tech, and telecom.

Customers don’t need historical data to use Lookout for Metrics. Customers only need their continuous data to get started as Lookout for Metrics learns from customer’s live data and will start showing results after an initialization period (when the models are learning from the data). The length of the initialization period will depend on the frequency of data, for example 5 min, 10 min, 1 hour, every 24 hours.

Lookout for Metrics uses data in a CSV or JSON format.

Amazon Lookout for Metrics easily connects to popular data stores such as Amazon Simple Storage (S3), Amazon Redshift, AWS CloudWatch, Amazon RDS (all supported database engines), and commonly use SaaS applications including Salesforce, Marketo, Zendesk, ServiceNow, Infor Nexus, Google Analytics, Dynatrace, Datadog, Amplitude, Veeva Systems, Singular, Amazon Simple Notification Service (SNS) AWS Lambda, Slack, PagerDuty, Datadog, Webhooks.

Lookout for Metrics is available in US East (N. Virginia), US West (Oregon), US East (Ohio), EU (Ireland), EU (Frankfurt), EU (Stockholm), Asia Pacific (Tokyo), Asia Pacific (Singapore), Asia Pacific (Sydney).

With Amazon Lookout for Metrics, you only pay for what you use. There is no up-front commitment or minimum fee. Amazon Lookout for Metrics automatically analyzes the metrics you care about, detects anomalies in those metrics, sends automated alerts, and can help you determine the severity of anomalies and their root cause. Please refer to the pricing page for more details and examples. 

You can get started here.