Example: Using with azure-functions-scaffold¶

This example shows how azure-functions-logging integrates with projects generated by azure-functions-scaffold.

Overview¶

Every project created with afs new includes azure-functions-logging as a default dependency and generates app/core/logging.py with a ready-to-use configuration. No extra flags or setup steps are required.

Generated Project Layout¶

my-api/
  app/
    core/
      logging.py      ← configure_logging() + module logger
    functions/
      http.py
  function_app.py     ← calls configure_logging() at startup
  pyproject.toml      ← azure-functions-logging listed under [project.dependencies]

Generated Files¶

`app/core/logging.py`¶

from __future__ import annotations

from azure_functions_logging import get_logger, setup_logging


def configure_logging() -> None:
    """Configure structured logging for the Azure Functions application."""
    setup_logging(format="json")


logger = get_logger("my-api")

`function_app.py` (startup call)¶

from __future__ import annotations

import azure.functions as func

from app.core.logging import configure_logging
from app.functions.http import http_blueprint

configure_logging()

app = func.FunctionApp()
app.register_functions(http_blueprint)

Using the Logger in Function Modules¶

Import the pre-created logger from app.core.logging in any function module.

from __future__ import annotations

import azure.functions as func

from app.core.logging import logger

http_blueprint = func.Blueprint()


@http_blueprint.route(route="hello", methods=["GET"], auth_level=func.AuthLevel.ANONYMOUS)
def hello(req: func.HttpRequest) -> func.HttpResponse:
    name = req.params.get("name", "world")
    logger.info("hello invoked", name=name)
    return func.HttpResponse(f"Hello, {name}!")

Injecting Invocation Context¶

Pass the context parameter to inject_context() so that invocation ID and function name are attached to every log record automatically.

from __future__ import annotations

import azure.functions as func

from azure_functions_logging import inject_context

from app.core.logging import logger

http_blueprint = func.Blueprint()


@http_blueprint.route(route="hello", methods=["GET"], auth_level=func.AuthLevel.ANONYMOUS)
def hello(req: func.HttpRequest, context: func.Context) -> func.HttpResponse:
    inject_context(context)
    name = req.params.get("name", "world")
    logger.info("hello invoked", name=name)
    return func.HttpResponse(f"Hello, {name}!")

JSON output in Azure Monitor will include invocation_id and function_name fields.

Customising Log Level per Environment¶

Override the level at startup without touching the generated file.

# function_app.py
import logging

from azure_functions_logging import setup_logging

setup_logging(level=logging.DEBUG, format="json")

Or keep format="color" locally and switch to "json" in production via an environment variable.

import logging
import os

from azure_functions_logging import setup_logging

log_format = os.environ.get("LOG_FORMAT", "json")
log_level = logging.DEBUG if os.environ.get("DEBUG") else logging.INFO

setup_logging(level=log_level, format=log_format)  # type: ignore[arg-type]

Scaffolding a New Project¶

pip install azure-functions-scaffold

# Create a project — logging is included by default
afs new my-api

# With additional integrations
afs new my-api --with-openapi --with-validation --with-doctor

Why This Pattern Works¶

setup_logging() runs once before any function registers, so all blueprints share the same configuration.
get_logger("my-api") gives a stable named logger scoped to the application, not the module path, which avoids naming conflicts across blueprints.
JSON format is the default because Azure Monitor ingests structured fields directly, enabling rich Kusto queries without text parsing.