Dynatrace offers Davis AI predictive forecasting that analyzes historical resource usage data to predict when cloud infrastructure — such as disk space, CPU, or memory — is likely to run out, allowing teams to act before an outage occurs rather than reacting after the fact.

Davis AI uses an AutoML (automated machine learning) approach that analyzes time series data stored in the Grail data lakehouse, detecting variance, seasonality, and trends to select the best forecasting model automatically. The system can predict resource consumption for thousands of individual components in parallel — for example, Dynatrace's own internal infrastructure tracks over 8,000 disks using this approach. For Kubernetes workloads, predictive AI can be combined with generative AI to automatically open pull requests on GitHub, suggesting scaling adjustments to manifest files for engineer review.