Amazon Lookout for Metrics is a absolutely managed service that makes use of machine studying (ML) to detect anomalies in nearly any time-series enterprise or operational metrics—reminiscent of income efficiency, buy transactions, and buyer acquisition and retention charges—with no ML expertise required. The service, which was launched in March 2021, predates a number of fashionable AWS choices which have anomaly detection, reminiscent of Amazon OpenSearch, Amazon CloudWatch, AWS Glue Data Quality, Amazon Redshift ML, and Amazon QuickSight.
After cautious consideration, we now have made the choice to finish help for Amazon Lookout for Metrics, efficient October 10, 2025. As well as, as of at the moment, new buyer sign-ups are now not accessible. Present clients will be capable to use the service as common till October 10, 2025, after we will finish help for Amazon Lookout for Metrics.
On this put up, we offer an summary of the alternate AWS companies that supply anomaly detection capabilities for clients to think about transitioning their workloads to.
AWS companies with anomaly detection capabilities
We suggest clients use Amazon OpenSearch, Amazon CloudWatch, Amazon Redshift ML, Amazon QuickSight, or AWS Glue Data Quality companies for their anomaly detection use instances as a substitute for Amazon Lookout for Metrics. These AWS companies provide usually accessible, ML-powered anomaly detection capabilities that can be utilized out of the field with out requiring any ML experience. Following is a short overview of every service.
Utilizing Amazon OpenSearch for anomaly detection
Amazon OpenSearch Service contains a extremely performant, built-in anomaly detection engine that permits the real-time identification of anomalies in streaming knowledge in addition to in historic knowledge. You may pair anomaly detection with built-in alerting in OpenSearch to ship notifications when there may be an anomaly. To start out utilizing OpenSearch for anomaly detection you first should index your data into OpenSearch, from there you may enable anomaly detection in OpenSearch Dashboards. To be taught extra, see the documentation.
Utilizing Amazon CloudWatch for anomaly detection
Amazon CloudWatch helps creating anomaly detectors on particular Amazon CloudWatch Log Teams by making use of statistical and ML algorithms to CloudWatch metrics. Anomaly detection alarms will be created based mostly on a metric’s anticipated worth. All these alarms don’t have a static threshold for figuring out alarm state. As a substitute, they examine the metric’s worth to the anticipated worth based mostly on the anomaly detection mannequin. To start out utilizing CloudWatch anomaly detection, you first should ingest data into CloudWatch after which enable anomaly detection on the log group.
Utilizing Amazon Redshift ML for anomaly detection
Amazon Redshift ML makes it simple to create, practice, and apply machine studying fashions utilizing acquainted SQL instructions in Amazon Redshift knowledge warehouses. Anomaly detection will be accomplished in your analytics knowledge by way of Redshift ML by utilizing the included XGBoost mannequin kind, native fashions, or distant fashions with Amazon SageMaker. With Redshift ML, you don’t must be a machine studying knowledgeable and also you pay solely for the coaching price of the SageMaker fashions. There aren’t any extra prices to utilizing Redshift ML for anomaly detection. To be taught extra, see the documentation.
Utilizing Amazon QuickSight for anomaly detection
Amazon QuickSight is a quick, cloud-powered, enterprise intelligence service that delivers insights to everybody within the group. As a completely managed service, QuickSight lets clients create and publish interactive dashboards that embody ML insights. QuickSight helps a extremely performant, built-in anomaly detection engine that makes use of confirmed Amazon expertise to constantly run ML-powered anomaly detection throughout hundreds of thousands of metrics to find hidden tendencies and outliers in clients’ knowledge. This software permits clients to get deep insights which are typically buried within the aggregates and never scalable with guide evaluation. With ML-powered anomaly detection, clients can discover outliers of their knowledge with out the necessity for guide evaluation, customized improvement, or ML area experience. To be taught extra, see the documentation.
Utilizing Amazon Glue Knowledge High quality for anomaly detection
Knowledge engineers and analysts can use AWS Glue Data Quality to measure and monitor their knowledge. AWS Glue Knowledge High quality makes use of a rule-based strategy that works properly for recognized knowledge patterns and gives ML-based suggestions that can assist you get began. You may evaluate the suggestions and increase guidelines from over 25 included knowledge high quality guidelines. To seize unanticipated, much less apparent knowledge patterns, you may allow anomaly detection. To make use of this function, you may write guidelines or analyzers after which activate anomaly detection in AWS Glue ETL. AWS Glue Knowledge High quality collects statistics for columns laid out in guidelines and analyzers, applies ML algorithms to detect anomalies, and generates visible observations explaining the detected points. Clients can use really useful guidelines to seize the anomalous patterns and supply suggestions to tune the ML mannequin for extra correct detection. To be taught extra, see the blog post, watch the introductory video, or see the documentation.
Utilizing Amazon SageMaker Canvas for anomaly detection (a beta function)
The Amazon SageMaker Canvas crew plans to offer help for anomaly detection use instances in Amazon SageMaker Canvas. We’ve created an AWS CloudFormation template-based resolution to provide clients early entry to the underlying anomaly detection function. Clients can use the CloudFormation template to convey up an utility stack that receives time-series knowledge from an Amazon Managed Streaming for Apache Kafka (Amazon MSK) streaming supply and performs near-real-time anomaly detection within the streaming knowledge. To be taught extra in regards to the beta providing, see Anomaly detection in streaming time series data with online learning using Amazon Managed Service for Apache Flink.
Continuously requested questions
- What’s the cutoff level for present clients?
We created an permit checklist of account IDs which have used Amazon Lookout for Metrics within the final 30 days and have lively Amazon Lookout for Metrics sources, together with detectors, throughout the service. If you’re an present buyer and are having difficulties utilizing the service, please attain out to us by way of AWS Customer Support for assist.
- How will entry change earlier than the sundown date?
Present clients can do all of the issues they might beforehand. The one change is that non-current clients can not create any new sources in Amazon Lookout for Metrics.
- What occurs to my Amazon Lookout for Metrics sources after the sundown date?
After October 10, 2025, all references to AWS Lookout for Metrics fashions and sources will probably be deleted from Amazon Lookout for Metrics. You won’t be able to find or entry Amazon Lookout for Metrics out of your AWS Administration Console and functions that decision the Amazon Lookout for Metrics API will now not work.
- Will I be billed for Amazon Lookout for Metrics sources remaining in my account after October 10, 2025?
Assets created by Amazon Lookout for Metrics internally will probably be deleted after October 10, 2025. Clients will probably be accountable for deleting the enter knowledge sources created by them, reminiscent of Amazon Easy Storage Service (Amazon S3) buckets, Amazon Redshift clusters, and so forth.
- How do I delete my Amazon Lookout for Metrics sources?
- How can I export anomalies knowledge earlier than deleting the sources?
Anomalies knowledge for every measure will be downloaded for a detector by utilizing the Amazon Lookout for Metrics APIs for a selected detector. Exporting Anomalies explains how to connect with a detector, question for anomalies, and obtain them right into a format for later use.
Conclusion
On this weblog put up, we now have outlined strategies to create anomaly detectors utilizing alternates reminiscent of Amazon OpenSearch, Amazon CloudWatch, and a CloudFormation template-based resolution.
Useful resource hyperlinks:
In regards to the Writer
Nirmal Kumar is Sr. Product Supervisor for the Amazon SageMaker service. Dedicated to broadening entry to AI/ML, he steers the event of no-code and low-code ML options. Outdoors work, he enjoys travelling and studying non-fiction.