A lightweight cloud engineering project that detects unusual infrastructure spend and sends explainable alerts before monthly costs drift out of control.
This item is used to test the ADE Showcase experience. It is not a real member submission and it does not carry ADE verification.
A fictional SaaS team noticed cost spikes only after invoices arrived. They needed an inexpensive way to compare recent spend against expected patterns and explain which service or environment caused a deviation.
The solution ingested daily billing exports, calculated rolling baselines, applied configurable anomaly thresholds and generated alerts with service-level context. The prototype used Python for processing, SQL for historical queries and a simple scheduled cloud function for automation.
The demo flagged synthetic cost spikes within one daily cycle and attached enough context for an engineer to investigate quickly. The project is presented as an engineering case study with architecture, assumptions and failure considerations.
Billing export, processing and alert flow
Baseline and threshold approach
Service, deviation and investigation context
False positives, missing data and scaling considerations