# Elastic Compute Autoscaling > Depth-first rabbithole dossier for Elastic Compute Autoscaling; source-anchored research pack. Parent: [MongoDB Atlas Cost Optimization](https://llms-explorer.com/tree/mongodb-atlas-cost-optimization/) · 2 facets · 2 facts · page: https://llms-explorer.com/tree/elastic-compute-autoscaling/ ## Problems, failure modes and limitations - [low] An error is raised if Atlas cannot start compute auto-scaling for operational nodes because the configured storage size is not supported by the target tier. Sources: https://www.mongodb.com/docs/atlas/reference/alert-conditions.md — source: `~/.global-ai-hub/research/.pipeline/mongodb-2026-10-14/frb/elastic-compute-autoscaling-aa9a52af7f/rabbithole-synthesis.md#troubleshooting` ## Facts and statements - [low] Atlas compute auto-scaling moves a dedicated cluster between a minimum and maximum tier based on sustained CPU and memory usage, allowing clusters with daily or weekly load cycles to run on a smaller tier outside their peaks. Sources: https://www.pointfive.co/efficiency-hub/cloud-services/mongodb-atlas — source: `~/.global-ai-hub/research/.pipeline/mongodb-2026-10-14/frb/elastic-compute-autoscaling-aa9a52af7f/rabbithole-synthesis.md#core` ## Related concepts - Compute — is a part of Elastic Compute Autoscaling