Queue Storage Trigger and Output Binding¶

Queue Storage is the simplest built-in background work primitive in Azure Functions. It is best for decoupling HTTP ingress from asynchronous processing when you need cheap at-least-once delivery and do not require ordering or rich broker features.

Trigger¶

Use @app.queue_trigger(...) to process a queue message.

Key parameters: queue_name, connection, and arg_name.

import azure.functions as func

app = func.FunctionApp()

@app.queue_trigger(arg_name="msg", queue_name="incoming-jobs", connection="AzureWebJobsStorage")
def queue_example(msg: func.QueueMessage) -> None:
    print(msg.get_body().decode("utf-8"))

Input Binding¶

Queue-triggered handlers often enrich a message with a Blob or SQL lookup.

import azure.functions as func

app = func.FunctionApp()

@app.queue_trigger(arg_name="msg", queue_name="images", connection="AzureWebJobsStorage")
@app.blob_input(arg_name="blob_data", path="uploads/{queueTrigger}", connection="AzureWebJobsStorage")
def queue_with_blob(msg: func.QueueMessage, blob_data: bytes) -> None:
    print(f"bytes={len(blob_data)}")

Output Binding¶

Use a queue output binding to chain work to another queue.

import json
import azure.functions as func

app = func.FunctionApp()

@app.queue_trigger(arg_name="msg", queue_name="incoming-jobs", connection="AzureWebJobsStorage")
@app.queue_output(arg_name="out_msg", queue_name="processed-jobs", connection="AzureWebJobsStorage")
def queue_fanout(msg: func.QueueMessage, out_msg: func.Out[str]) -> None:
    out_msg.set(json.dumps({"source": msg.id, "status": "done"}))

Configuration¶

Queue polling and concurrency live under extensions.queues in host.json:

{
  "version": "2.0",
  "extensions": {
    "queues": {
      "batchSize": 16,
      "newBatchThreshold": 8,
      "maxDequeueCount": 5,
      "visibilityTimeout": "00:00:30"
    }
  }
}

local.settings.json almost always includes AzureWebJobsStorage; identity-based storage connections can replace raw connection strings in deployed environments.

Scaling Behavior¶

The scale controller looks at queue depth and message age, then adds instances as backlog grows. Each instance polls in batches and hides messages during processing with the visibility timeout. If processing fails or exceeds visibility, the message can reappear and be retried on the same or another instance.

Common Pitfalls¶

Queue Storage is not FIFO once retries, invisibility timeouts, and multiple workers are involved.
If processing takes longer than visibilityTimeout, duplicate work can happen unless the handler is idempotent.
Poison messages move only after maxDequeueCount; watch the poison queue instead of assuming failures disappear.

https://learn.microsoft.com/azure/azure-functions/functions-bindings-storage-queue-trigger?pivots=programming-language-python&tabs=python-v2
https://learn.microsoft.com/azure/azure-functions/functions-bindings-storage-queue-output?pivots=programming-language-python&tabs=python-v2