Manage Azure Queue Storage with Python

Azure Queue Storage SDK for Python is a client for simple, cost-effective message queuing for asynchronous communication, via a QueueServiceClient (account-level: create/delete/list queues) and a QueueClient (per-queue operations), both authenticating with DefaultAzureCredential, with a mirrored async client for high-throughput use.

Sending covers a plain string message, a message with options (visibility_timeout to delay when it becomes visible, time_to_live to expire it), or a JSON payload serialized to a string; binary content is supported via BinaryBase64EncodePolicy/BinaryBase64DecodePolicy configured on the client so raw bytes can be sent and read back directly. Receiving pulls a page of messages (messages_per_page) with a chosen visibility timeout (how long they stay hidden from other receivers while you process them), exposing each message's ID, content, and dequeue_count - a rising count signals a poison message that keeps failing processing. Messages must be explicitly deleted after successful processing; an unhandled exception leaves the message to reappear once its visibility timeout expires, giving automatic at-least-once retry. peek_messages() reads messages without affecting their visibility, useful for monitoring the queue without competing with real consumers. update_message() can extend a message's visibility timeout mid-processing or replace its content while keeping it in place. The whole queue can be cleared in one call, and queue-level properties (an approximate message count) and custom metadata can be read and set.

Use it for simple, high-volume asynchronous work queues where you just need reliable message delivery with basic visibility-timeout-based retry. It explicitly recommends Service Bus instead when you need more advanced messaging features - sessions, topics/subscriptions, or transactions - that this simpler queue service doesn't provide.

Input is a string, JSON-serialized, or binary message sent to a named queue. Output is delivered messages you read, process, and explicitly delete (or let time out for retry), plus queue metadata and an approximate message count. Install it with:

pip install azure-storage-queue azure-identity

Its own best practices: always delete messages after processing to prevent reprocessing, set a visibility timeout that matches your actual processing time, watch dequeue_count to detect poison messages, use the async client for high-throughput scenarios, use peek_messages for monitoring without disturbing the queue, set time_to_live to avoid stale messages piling up, and reach for Service Bus instead when you need sessions or topics.

It's the azure-storage-queue package paired with azure-identity for DefaultAzureCredential authentication against an Azure Storage account's queue endpoint.

Python developers who need simple, cost-effective asynchronous message queuing with basic retry semantics, and who don't yet need Service Bus's sessions, topics, or transactional features.