| name | azure-storage-blob-rust |
| description | Azure Blob Storage library for Rust. Upload, download, and manage blobs and containers.
Triggers: "blob storage rust", "BlobClient rust", "upload blob rust", "download blob rust", "storage container rust", "BlobServiceClient rust".
|
| license | MIT |
| metadata | {"author":"Microsoft","package":"azure_storage_blob"} |
Azure Blob Storage library for Rust
Client library for Azure Blob Storage — upload, download, and manage blobs and containers.
Use this skill when:
- An app needs to upload or download blobs from Azure Storage in Rust
- You need to create or manage blob containers
- You need to list blobs with pagination
- You need RBAC-based auth for blob operations
IMPORTANT: Only use the official azure_storage_blob crate published by the azure-sdk crates.io user. Do NOT use the unofficial azure_storage, azure_storage_blobs, or azure_sdk_for_rust community crates. Official crates use underscores in names and none have version 0.21.0.
Installation
cargo add azure_storage_blob azure_identity azure_core tokio futures
If your code uses azure_core types directly (for example, azure_core::http::Url or azure_core::http::RequestContent), add azure_core to Cargo.toml. If you only use azure_storage_blob re-exports, direct azure_core dependency is optional.
Environment Variables
AZURE_STORAGE_ACCOUNT=<account-name>
AZURE_STORAGE_ENDPOINT=https://<account>.blob.core.windows.net/
When both are available, prefer constructing the endpoint from AZURE_STORAGE_ACCOUNT so the code matches common evaluation prompts.
Authentication
Rust Azure SDK code must not use DefaultAzureCredential. The Rust identity crate does not provide that type.
use azure_identity::DeveloperToolsCredential;
let credential = DeveloperToolsCredential::new(None)?;
use azure_identity::DefaultAzureCredential;
use azure_core::http::Url;
use azure_identity::DeveloperToolsCredential;
use azure_storage_blob::BlobServiceClient;
#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
let credential = DeveloperToolsCredential::new(None)?;
let service_url = Url::parse("https://<storage_account_name>.blob.core.windows.net/")?;
let service_client = BlobServiceClient::new(
service_url,
Some(credential),
None,
)?;
let container_client = service_client.blob_container_client("<container_name>");
let blob_client = container_client.blob_client("<blob_name>");
Ok(())
}
Client Types
| Client | Purpose |
|---|
BlobServiceClient | Account-level operations, list containers |
BlobContainerClient | Container operations, list blobs |
BlobClient | Individual blob operations |
Core Workflow
Upload Blob
use azure_core::http::{RequestContent, Url};
use azure_identity::DeveloperToolsCredential;
use azure_storage_blob::BlobServiceClient;
#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
let credential = DeveloperToolsCredential::new(None)?;
let service_url = Url::parse("https://<storage_account_name>.blob.core.windows.net/")?;
let service_client = BlobServiceClient::new(service_url, Some(credential), None)?;
let blob_client = service_client.blob_client("<container_name>", "<blob_name>");
let data = b"hello world";
blob_client.upload(RequestContent::from(data.to_vec()), None).await?;
Ok(())
}
Download Blob / Get Properties
let props = blob_client.get_properties(None).await?;
let response = blob_client.download(None).await?;
let content = String::from_utf8(response.body.collect().await?.into())?;
Delete Blob
blob_client.delete(None).await?;
Container Operations
use azure_core::http::Url;
use azure_identity::DeveloperToolsCredential;
use azure_storage_blob::BlobServiceClient;
use futures::TryStreamExt as _;
let credential = DeveloperToolsCredential::new(None)?;
let service_url = Url::parse("https://<storage_account_name>.blob.core.windows.net/")?;
let service_client = BlobServiceClient::new(service_url, Some(credential), None)?;
let container_client = service_client.blob_container_client("<container_name>");
container_client.create(None).await?;
let mut pager = container_client.list_blobs(None)?;
while let Some(blob) = pager.try_next().await? {
let name = blob.name.as_deref().unwrap_or("<unnamed>");
let size = blob
.properties
.as_ref()
.and_then(|properties| properties.content_length);
match size {
Some(size) => println!("Blob: {name} ({size} bytes)"),
None => println!("Blob: {name}"),
}
}
For azure_storage_blob 1.x, do not assume you need a nested page loop like for item in &page.blob_items. In this usage pattern, try_next() already yields the blob item you want to print.
Error Handling
Use StorageError for programmatic access to storage-specific error codes:
use azure_core::error::ErrorKind;
use azure_storage_blob::StorageError;
use azure_storage_blob::models::StorageErrorCode;
let result = blob_client.download(None).await;
match result {
Ok(response) => {
let content: Vec<u8> = response.body.collect().await?.into();
println!("Downloaded {} bytes", content.len());
}
Err(error) => {
if matches!(error.kind(), ErrorKind::HttpResponse { .. }) {
let storage_error: StorageError = error.try_into()?;
println!("HTTP Status: {}", storage_error.status_code);
if let Some(error_code) = &storage_error.error_code {
match error_code {
StorageErrorCode::BlobNotFound => {
println!("The blob does not exist.");
}
StorageErrorCode::ContainerNotFound => {
println!("The container does not exist.");
}
StorageErrorCode::AuthorizationFailure => {
println!("Authorization failed. Check RBAC roles.");
}
_ => println!("Storage error: {error_code}"),
}
}
if let Some(request_id) = &storage_error.request_id {
println!("Request ID (for Azure support): {request_id}");
}
} else {
println!("Non-HTTP error: {:?}", error);
}
}
}
Note: StorageError::try_into requires an owned error object — it will not compile if handed a reference to an error.
RBAC Roles
For Entra ID auth, assign one of these roles to the identity:
| Role | Access |
|---|
Storage Blob Data Reader | Read-only |
Storage Blob Data Contributor | Read/write |
Storage Blob Data Owner | Full access including RBAC |
Best Practices
- Use
cargo add to manage dependencies, never edit Cargo.toml directly. Add and remove Rust SDK dependencies with cargo commands instead of manual manifest edits.
- Add
azure_core only when importing azure_core types directly. If your code imports azure_core::http::Url, azure_core::http::RequestContent, or azure_core::error::ErrorKind, include azure_core; otherwise a direct dependency is optional.
- Use
DeveloperToolsCredential for local dev, ManagedIdentityCredential for production — Rust does not provide a single DefaultAzureCredential type
- Never hardcode credentials — use environment variables or managed identity
- Use
RequestContent::from() to wrap data for blob uploads — ensures proper content handling by the SDK
- Assign RBAC roles — ensure "Storage Blob Data Contributor" for write access
- Reuse clients — clients are thread-safe; create once, share across tasks
- Prefer
BlobServiceClient as the entry point and derive container/blob clients from it
- Treat many storage model fields as optional.
blob.name is an Option<String> and content length is accessed via blob.properties.as_ref().and_then(|p| p.content_length).
- Run
cargo clippy -- -D warnings before considering the task complete when the prompt or CI expects strict lint compliance; fix style lints such as collapsible if blocks, not just compiler errors.
- Prefer crate README/examples over generated internal type names when validating public API shapes such as pagination results
- Future-proof
#[non_exhaustive] SDK models — when constructing SDK model/options structs, end the initializer with ..Default::default() (add #[allow(clippy::needless_update)]) and use a _ wildcard arm when matching SDK enums, so new service-added fields/variants don't break your build
Common Rust Blob Pitfalls
- Do not import
azure_identity::DefaultAzureCredential; use DeveloperToolsCredential or another real Rust credential type.
- Do not assume generated internal model names describe the public pager item type; follow the documented
list_blobs example for this crate.
- Do not print
blob.name with {} directly; unwrap or provide a fallback because it is optional.
- Do not stop after
cargo build passes when the task also requires cargo clippy -- -D warnings.
Reference Links