Amazon Lookout for Metrics uses machine learning (ML) to analyze nearly any time-series business or operational metric (such as revenue performance, purchase transactions, customer acquisition and retention rates, etc.) without any ML experience. A fully managed service that detects anomalies. . The service launched in March 2021 and predates several popular AWS products with anomaly detection capabilities, including Amazon OpenSearch, Amazon CloudWatch, AWS Glue Data Quality, Amazon Redshift ML, and Amazon QuickSight. .
After careful consideration, we have decided to end support for Amazon Lookout for Metrics on October 10, 2025. Additionally, as of today, sign-up for new customers is no longer available. Existing customers can continue to use the service as usual until October 10, 2025, when support for Amazon Lookout for Metrics ends.
This post provides an overview of alternative AWS services that provide anomaly detection capabilities for customers considering migrating their workloads.
AWS services with anomaly detection capabilities
For anomaly detection use cases, we recommend using Amazon OpenSearch, Amazon CloudWatch, Amazon Redshift ML, Amazon QuickSight, or the AWS Glue Data Quality service instead of Amazon Lookout for Metrics. These AWS services provide generally available ML-powered anomaly detection capabilities that you can use out of the box without requiring ML expertise. Below is a brief overview of each service.
Using Amazon OpenSearch for anomaly detection
Amazon OpenSearch Service has a high-performance, integrated anomaly detection engine that can identify anomalies in streaming and historical data in real time. You can combine anomaly detection with OpenSearch’s built-in alerts to send notifications when an anomaly occurs. To start using OpenSearch for anomaly detection, you must first index your data into OpenSearch. From there, you can enable anomaly detection in your OpenSearch dashboard. See the documentation for more information.
Anomaly detection using Amazon CloudWatch
Amazon CloudWatch supports creating anomaly detectors on specific Amazon CloudWatch log groups by applying statistical and ML algorithms to CloudWatch metrics. Anomaly detection alarms can be created based on the expected value of a metric. These types of alarms do not have static thresholds to determine the state of the alarm. Instead, it compares the value of the metric to the expected value based on an anomaly detection model. To start using CloudWatch anomaly detection, you must first ingest data into CloudWatch and then enable anomaly detection on your log group.
Use Amazon Redshift ML for anomaly detection
Amazon Redshift ML allows you to easily create, train, and apply machine learning models using familiar SQL commands in your Amazon Redshift data warehouse. Anomaly detection can be performed on analytical data through Redshift ML using the included XGBoost model type, local models, or remote models in Amazon SageMaker. With Redshift ML, you don’t need to be a machine learning expert and only pay for the training costs of your SageMaker models. There is no additional cost to use Redshift ML for anomaly detection. See the documentation for more information.
Using Amazon QuickSight for anomaly detection
Amazon QuickSight is a fast, cloud-powered business intelligence service that delivers insights to everyone in your organization. As a fully managed service, QuickSight allows customers to create and publish interactive dashboards containing ML insights. QuickSight supports a high-performance, unified anomaly detection engine using proven Amazon technology to continuously perform ML-driven anomaly detection across millions of metrics to uncover hidden trends and outliers in customer data. I will. This tool allows customers to gain deep insights that are often buried in aggregations and cannot be extended through manual analysis. ML-powered anomaly detection allows customers to find outliers in their data without the need for manual analysis, custom development, or ML domain expertise. See the documentation for more information.
Using Amazon Glue Data Quality for anomaly detection
Data engineers and analysts can use AWS Glue Data Quality to measure and monitor their data. AWS Glue Data Quality uses a rules-based approach that fits your known data patterns and provides ML-based recommendations to help you get started. Review recommendations and enhance your rules from over 25 included data quality rules. Enable anomaly detection to capture unexpected, less obvious data patterns. To use this feature, create a rule or analyzer and enable anomaly detection in AWS Glue ETL. AWS Glue Data Quality collects statistics on columns specified by rules and analyzers, applies ML algorithms to detect anomalies, and produces visual observations that describe detected issues. Customers can use recommended rules to capture anomalous patterns and provide feedback to tune the ML model for more accurate detection. For more information, see our blog post, introductory video, or documentation.
Anomaly Detection with Amazon SageMaker Canvas (Beta Feature)
The Amazon SageMaker Canvas team plans to provide support for anomaly detection use cases in Amazon SageMaker Canvas. We created an AWS CloudFormation template-based solution to give our customers early access to the underlying anomaly detection capabilities. Customers can use CloudFormation templates to launch application stacks that receive time series data from Amazon Managed Streaming for Apache Kafka (Amazon MSK) streaming sources and perform near real-time anomaly detection within the streaming data. For more information about the beta, see Anomaly Detection in Streaming Time Series Data with Online Learning Using Amazon Managed Service for Apache Flink.
FAQ
- What is the cut-off point for current customers?
You have created an allowlist of account IDs that have used Amazon Lookout for Metrics in the past 30 days and have active Amazon Lookout for Metrics resources that include detectors within the service. If you are an existing customer and are having trouble using the service, please seek assistance through AWS Customer Support.
- how Will access change before sunset?
Today’s customers can do everything they could before. The only change is that customers other than the current customer cannot create new resources in Amazon Lookout for Metrics.
- What happens to my Amazon Lookout for Metrics resources after sunset?
Starting October 10, 2025, all references to AWS Lookout for Metrics models and resources will be removed from Amazon Lookout for Metrics. You can no longer discover or access Amazon Lookout for Metrics from the AWS Management Console, and applications that call the Amazon Lookout for Metrics API will no longer function.
- Will I be charged for any Amazon Lookout for Metrics resources that remain in my account after October 10, 2025?
Resources created internally by Amazon Lookout for Metrics will be deleted after October 10, 2025. You are responsible for deleting any input data sources that you created, such as Amazon Simple Storage Service (Amazon S3) buckets or Amazon Redshift clusters. .
- How do I delete my Amazon Lookout for Metrics resources?
- How do I export anomaly data before deleting a resource?
Anomaly data for each measure can be downloaded for a detector using the specific detector’s Amazon Lookout for Metrics API. Exporting Anomalies describes how to connect to a detector, query for anomalies, and download them to a format that can be used later.
conclusion
In this blog post, we outlined how to create anomaly detection functionality using alternatives such as Amazon OpenSearch, Amazon CloudWatch, and CloudFormation template-based solutions.
Resource link:
About the author
Nirmal Kumar I am a senior product manager for the Amazon SageMaker service. He is committed to expanding access to AI/ML and leads the development of no-code and low-code ML solutions. Outside of work, I enjoy traveling and reading nonfiction.
