Build a real-time, low-code anomaly detection pipeline for time series data using Amazon Aurora, Amazon Redshift ML, and Amazon SageMaker

This article explains how to build a low-code, real-time anomaly detection pipeline for time series data using Amazon Aurora, Amazon Redshift ML, and Amazon SageMaker.

Specifically, the article covers:

• Overview of the solution architecture
• Prerequisites (Aurora instance, Redshift data warehouse with Redshift ML)
• Source data overview (turbofan engine sensor data)
• Creating Aurora source database and populating data
• Setting up zero-ETL integration between Aurora and Redshift
• Creating materialized view in Redshift
• Anomaly detection using SageMaker Random Cut Forest model
• Integrating the SageMaker model with Redshift ML
• Making predictions on live data using the integrated model
• Cleaning up resources