What it is

Atlas Stream Processing (ASP) Vertical Autoscaling is a capacity management feature that dynamically aligns stream processor compute resources with real-time CPU & memory utilization. It automatically optimizes performance and operational efficiency by scaling processors up during high load and down during lulls, operating entirely within customer-defined cost governance boundaries (minTier and maxTier) to prevent unexpected cloud infrastructure spend.

Who it’s for

This feature is designed for engineering leads, platform architects, DevOps managers, and financial operations (FinOps) owners who oversee streaming data pipelines. It is tailored for organizations looking to eliminate manual resource provisioning and reduce cloud infrastructure overhead while ensuring high availability for critical, variable-volume data streaming workloads.

Why it matters

It eliminates the business risks and hidden costs of manual infrastructure management—specifically preventing middle-of-the-night outages from sudden traffic spikes while curbing excessive spending on over-provisioned capacity. By automating workload right-sizing within strict floor and ceiling bounds, enterprises maintain optimal service reliability, improve operational efficiency, and keep data processing costs aligned with actual business activity.

How to get started

Teams can activate Vertical Autoscaling across stream processors via the Atlas UI, Admin API, mongosh CLI, or Terraform infrastructure-as-code scripts by defining optional minTier and maxTier boundaries. Once enabled, MongoDB automatically manages the dynamic scaling lifecycle according to internal resource utilization metrics, emitting audit-ready updates to activity feeds and monitoring dashboards without requiring ongoing administrative intervention.