Hi there,
Is my understanding correct, that the multivariate anomaly detection API can only detect anomalies on a batch of timeseries data, and might therefore not be very suitable for real-time anomaly detection (i.e. streaming anomaly detection)? Are there potential alternatives?