The optimized autoscale feature has been in place and using reactive logic. It has been helping ADX users by adjusting the cluster size when there is an increase in resources due to ingestion or query load. The optimized autoscale feature has been further improved with predictive logic. This logic monitors the same metrics as the reactive logic and over time builds up the cluster usage pattern and uses this to forecast and plan the size of the cluster. The reactive logic is still used to ensure any forecast anomalies or usage pattern changes are still autoscaled appropriately.
There are a few cost-related updates but also updates to visuals, data ingestion, and more.