Repository navigation
Azure Workload Dataset v1
Curated subset of Microsoft Azure Public Dataset V2 for the Cloud Workload Forecasting assignment.
Contents
- 10,000 Azure VMs
- 1,340,000 CPU measurements
- 134 measurements per VM
- 5-minute sampling interval
- VM-level CPU readings and corresponding VM metadata
- Approximately 11.1 hours of continuous observations per VM
The dataset was deterministically sampled from complete VM traces in the first three chronological CPU-reading shards of Azure Public Dataset V2.
See DATA_CARD.md inside the release archive for the dataset schema, selection methodology, validation details, and source information.
Usage
Download data-v1.tar.gz, extract it, and verify the included SHA256 checksums before using the dataset.
This release is versioned and should be treated as immutable.