Identifying suitable metrics for monitoring
You have successfully framed your ML project as an anomaly or outlier detection problem and you may have collected some historical time series datasets. Is Amazon Lookout for Metrics a good candidate to deliver your desired insights? Let's review some considerations that will help you understand whether Amazon Lookout for Metrics is suitable for your anomaly detection scenario, namely, the following:
- Dataset requirements
- Use case requirements
Dataset requirements
Unlike many other AI services (within AWS or not), Amazon Lookout for Metrics can get started without any data points: you won't receive any detected anomalies until there is enough information to perform a cold start detection (for instance, if your data has daily intervals, you will need to wait for at least 14 days to start receiving inference results from Amazon Lookout for Metrics) but you can start building your pipeline and your applications...