Note
This page applies to both Atlas Infinite and Atlas Core.
Atlas uses auto-scaling to help you optimize resource utilization and cost. Cluster tier auto-scaling adjusts your cluster tier on both database editions. How cluster storage grows depends on the database edition.
Atlas uses two mechanisms to auto-scale the cluster tier. Reactive auto-scaling responds to your cluster's current resource usage and applies to every eligible cluster. Predictive auto-scaling uses historical usage patterns to scale ahead of a forecasted spike. Predictive auto-scaling extends reactive auto-scaling and falls back to it when a spike isn't cyclical or predictable.
Compare Auto-Scaling on Each Database Edition
This section summarizes how auto-scaling differs between the two database editions.
Cluster tier. Both database editions use reactive and predictive cluster tier auto-scaling. To learn more, see Compute Auto-Scaling on Atlas Core and Compute Auto-Scaling on Atlas Infinite.
Cluster storage. On an Atlas Core cluster, Atlas increases your storage capacity when the cluster uses 90% of its disk capacity, and you configure or opt out of it. On an Atlas Infinite cluster, storage grows with your data automatically. To learn more, see Auto-Scaling for Cluster Storage.
The oplog also affects auto-scaling differently on each edition:
On an Atlas Core cluster, Atlas delays an auto-scaling event until the oplog window covers the estimated scaling time. To keep the window from blocking scaling, set the minimum oplog retention window.
On an Atlas Infinite cluster, the minimum oplog retention window doesn't affect auto-scaling. To learn more, see Oplog on Atlas Infinite clusters.