Recent alerts

When AWS does raise a price

AWS list prices rarely go up, and the live demo never invents a price change, so the alerts above will not show one. These three cards are real On-Demand list-price increases from the AWS Price List, applied to synthetic Acme Corp spend and run through the same detection and explanation code on the date each change happened. The price facts are AWS's. Acme's quantities and dollar amounts are invented.

About this demo

The account above is synthetic. Acme Corp is not a real company and its spend is generated, not copied from any customer. A generator writes a new day of billing data for it every night, with anomalies rotating in, and the real CostPulse pipeline processes it exactly as it processes a customer's AWS Cost and Usage Report: the same baselines, the same detection rule and the same alert text.

A spike alert fires when a service's spend in one region is 20% or more above its 7-day rolling average (the default threshold, configurable per alert channel), and also above its average for the same weekday once there is enough history. The bar doubles on the 1st and 2nd of the month, when monthly charges land. Spikes of 50% or more are marked critical. Spend of $25 or more a day on a service and region with no spend in the previous 35 days alerts as new spend.

Opening an alert shows why it fired: which usage types moved, and a verdict on the cause. The verdict separates your usage, your rate, an AWS list-price change and a new charge, and says mixed or unknown when the data cannot single one out. That is the same explanation the Slack, Teams and email alert carried.