Updated 26 October 2026 · Published by Zoorik
Cloud cost anomaly detection is an automated check that learns your normal cloud spend and flags sudden changes. The cost anomaly meaning is narrow: a spike or drop your past spend doesn’t predict.
What is a cost anomaly in AWS and Azure?
In AWS, it’s spend that AWS Cost Anomaly Detection marks as unusual, using machine learning. Microsoft Cost Management finds anomalies per Azure subscription. Both can email you an alert.
- AWS allows for weekly or monthly patterns and natural growth.
- AWS checks about three times a day, but detection can take up to 24 hours.
- AWS needs 10 days of history for a newly used service.
- Azure compares each day’s total with a forecast from the last 60 days.
- Azure runs that check 36 hours after the day ends (UTC).
An example: a costlier gateway
In Microsoft’s 2021 example, a resource group cost about $4.56 a day.
On 30 June, a VpnGw1 charge on its virtual private network (VPN) gateway stopped, and a costlier VpnGw3 charge started. Daily cost rose to $20.68 that day and $30.22 from 1 July.
Why it matters for your cloud cost
- An alert flags a sudden jump within days. If it wasn’t planned, you catch the mistake early.
- Detection misses steady over-provisioning, such as a virtual machine sized above its load from day one. That cost is part of the normal it learns.
- Finding it takes a review of what you run against what it uses.
- The free assessment works out what your clusters and disks could save.
Read next: How to reduce cloud costs
Sources: Amazon Web Services, “Detecting unusual spend with AWS Cost Anomaly Detection” (opens in a new tab); Microsoft Learn, “Identify anomalies and unexpected changes in cost” (opens in a new tab); Microsoft Learn, “Anomaly management” (opens in a new tab).