Snapshots-based monitoring for data assets without time fields

You can create ML-based monitoring on tables that are not having any time fields. The ML models will continuously take snapshots of your data at every run to compute a time series based on which it trains and detects anomalies.

To do so, when creating an ML-based monitor, you must leave the Time Window box unchecked in the parameters.

📘

Training time

The model will require at least 10 data points to train and start detecting anomalies. This means that there won't be any anomaly detected during the first 10 runs of the monitor.

With this feature, users can monitor more than just the metadata of dimension tables with no time fields, such as a sudden increase of formatting issues in the customers emails field, or a sudden drop of the average lifetime value of your customers.