Performance and Scaling Basics¶
Understanding performance limits and scaling targets is essential for designing efficient Azure Storage solutions.
| Metric | Standard Account | Premium Block Blob | Premium File Share |
|---|---|---|---|
| Request Rate | Up to 40,000 requests/sec (default in listed regions); 20,000 requests/sec in other regions | Service-specific premium targets (see source) | Up to 102,400 (provisioned SSD) |
| Ingress | Up to 60 Gbps (default in listed regions); 25 Gbps in other regions | Service-specific premium targets (see source) | Service/account-level throughput targets apply |
| Egress | Up to 200 Gbps (default in listed regions); 50 Gbps in other regions | Service-specific premium targets (see source) | Service/account-level throughput targets apply |
| Capacity | 5 PiB per account (default) | Service-specific premium targets (see source) | Up to 256 TiB (provisioned v2) |
mermaid graph TD App[App Request] --> Th{Throttled?} Th -- Yes --> Client{Client Action} Client --> Parallel[Increase Parallelism] Client --> Retry[Exponential Backoff] Th -- No --> Partition[Partition Check] Partition --> Key[Optimize Partition Key] Key --> Success[High Throughput]
Note
Limits are region-dependent and workload-dependent. Higher capacity and ingress/egress limits can be requested through Azure Support. See the Microsoft Learn source for the current list of regions that receive the higher default targets.
Key Concepts¶
- Throughput: The amount of data transferred per second.
- IOPS: The number of input/output operations per second.
- Partitioning: Azure Storage uses a partition key to scale data across multiple servers.