Anomaly Detector
An AI service that helps you foresee problems before they occur
Boost the reliability of your business by detecting problems early
Easily embed time-series anomaly detection capabilities into your apps to help users identify problems quickly. Anomaly Detector ingests time-series data of all types and selects the best anomaly-detection algorithm for your data to ensure high accuracy. Detect spikes, dips and deviations from cyclic patterns, and trend changes through both univariate and multivariate APIs. Customise the service to detect any level of anomaly. Deploy the anomaly detection service where you need it – in the cloud or at the intelligent edge.
Powerful inference engine assesses your time-series dataset and automatically selects the right anomaly detection algorithm to maximise accuracy for your scenario.
Automatic detection eliminates the need for labelled training data to help you save time and stay focused on fixing problems as soon as they surface.
Customisable settings let you fine-tune sensitivity to potential anomalies based on the risk profile of your business.
Speed your time to insights
Fast-track your problem-solving with simple setup in the Azure portal and real-time anomaly detection systems. All it takes is three lines of code.
Identify multivariate anomalies
Use multivariate anomaly detection to evaluate multiple signals and the correlations between them to find sudden changes in data patterns before they affect your business.
Detect problems for virtually any scenario
There are many types of time-series data, and no one algorithm fits them all. Anomaly Detector assesses your time-series data set and automatically selects the best algorithm and the best anomaly-detection techniques from the model gallery. Use the service to ensure high accuracy for scenarios, including monitoring IoT device traffic, managing fraud and responding to changing markets.
Trusted by Microsoft Azure, Office, Windows and Bing
Monitor your product and service health and deliver reliable customer experiences using the same anomaly detection system and service that more than 200 Microsoft product teams rely on.
Airbus deployed Anomaly Detector, part of Cognitive Services, to monitor the condition of an aircraft and fix potential problems before they occur. The company developed a proof of concept for the aircraft-monitoring application using multivariant anomaly detection, loading telemetry data from multiple flights for analysis and model training.
GA Insights uses Anomaly Detector to monitor business metrics and alert customers about anomalies in data patterns, such as e-commerce conversion rates, sales or revenue before they effect their business.
TIBCO used Anomaly Detector and Azure Cognitive Services to develop its own container-based anomaly detection solution, which alerts case managers to sudden changes in data patterns, discovers root causes and provides suggested actions to reduce costly downtime of devices and equipment.
Industry-leading security for the enterprise
- Microsoft invests over USD 1 billion annually on cyber security research and development.
- We employ more than 3,500 security experts who are completely focused on securing your data and privacy.
- Azure has more certifications than any other cloud provider. View the comprehensive list.
Documentation and resources
Getting started
Read our documentation
Learn how it works
Watch this episode of the AI Show on Channel 9 for a guided walkthrough on setting up Anomaly Detector.
Get started in three steps
Create your Azure free account.
Sign in to the Azure portal.
Explore the documentation and quickstarts.
Stay updated on the latest Cognitive Services news
Frequently asked questions about Anomaly Detector
-
Please see regional availability of univariate anomaly detection feature. Multivariate anomaly detection (preview) feature is available in selected regions. Please check this document for details.
-
Anomaly Detector provides a 99.9 per cent service-level agreement (SLA).
-
Anomaly Detector is comprised of simple REST APIs with a code-first experience. It's the core engine of Metrics Advisor that detects anomalies in time-series data. It's best applied for ad-hoc data analysis, and it can be run in containers. Metrics Advisor has additional time-series monitoring features, with pipeline APIs and a built-in user interface for managing the service. It's designed for live-streaming data and AI analytics, and it supports deploying in Azure.
-
It can vary depending on the level of accuracy and speed desired for your scenario. Read the best practices guide for more details.