This SDK is no longer maintained or supported by Microsoft. Check out the Python Open Census SDK <https://github.com/census-instrumentation/opencensus-python>_ for Azure Monitor's latest Python investments. Azure Monitor only provides support when using the supported SDKs <https://docs.microsoft.com/en-us/azure/azure-monitor/app/platforms#unsupported-community-sdks>. We’re constantly assessing opportunities to expand our support for other languages, so follow our GitHub Announcements <https://github.com/microsoft/ApplicationInsights-Announcements/issues> page to receive the latest SDK news.

Application Insights for Python

.. image:: https://travis-ci.org/Microsoft/ApplicationInsights-Python.svg?branch=master :target: https://travis-ci.org/Microsoft/ApplicationInsights-Python

.. image:: https://badge.fury.io/py/applicationinsights.svg :target: http://badge.fury.io/py/applicationinsights

This project extends the Application Insights API surface to support Python. Application Insights <http://azure.microsoft.com/services/application-insights/>_ is a service that allows developers to keep their application available, performing and succeeding. This Python module will allow you to send telemetry of various kinds (event, trace, exception, etc.) to the Application Insights service where they can be visualized in the Azure Portal. A link to the Application Insights API documentation can be found here <https://microsoft.github.io/ApplicationInsights-Python/>_.

This project is not officially supported and not recommended for high load production use. The project is open source and welcomes contributions. Please refer to CONTRIBUTING.md <https://github.com/Microsoft/ApplicationInsights-Python/blob/develop/CONTRIBUTING.md>_ for details.

Requirements

Python >=2.7 and Python >=3.4 are currently supported by this module.

Installation

To install the latest release you can use pip <http://www.pip-installer.org/>_.

::

$ pip install applicationinsights

Documentation

Please see https://microsoft.github.io/ApplicationInsights-Python/ for full documentation.

Usage

Once installed, you can send telemetry to Application Insights. Here are a few samples.

**Note**: before you can send data to you will need an instrumentation key. Please see the `Getting an Application Insights Instrumentation Key <https://github.com/Microsoft/AppInsights-Home/wiki#getting-an-application-insights-instrumentation-key>`_ section for more information.

Sending a simple event telemetry item

.. code:: python

from applicationinsights import TelemetryClient
tc = TelemetryClient('<YOUR INSTRUMENTATION KEY GOES HERE>')
tc.track_event('Test event')
tc.flush()

Sending an event telemetry item with custom properties and measurements

.. code:: python

from applicationinsights import TelemetryClient
tc = TelemetryClient('<YOUR INSTRUMENTATION KEY GOES HERE>')
tc.track_event('Test event', { 'foo': 'bar' }, { 'baz': 42 })
tc.flush()

Sending a trace telemetry item with custom properties

.. code:: python

from applicationinsights import TelemetryClient
tc = TelemetryClient('<YOUR INSTRUMENTATION KEY GOES HERE>')
tc.track_trace('Test trace', { 'foo': 'bar' })
tc.flush()

Sending a metric telemetry item

.. code:: python

from applicationinsights import TelemetryClient
tc = TelemetryClient('<YOUR INSTRUMENTATION KEY GOES HERE>')
tc.track_metric('My Metric', 42)
tc.flush()

Sending an exception telemetry item with custom properties and measurements

.. code:: python

import sys
from applicationinsights import TelemetryClient
tc = TelemetryClient('<YOUR INSTRUMENTATION KEY GOES HERE>')
try:
    raise Exception('blah')
except:
    tc.track_exception()

try:
    raise Exception("blah")
except:
    tc.track_exception(*sys.exc_info(), properties={ 'foo': 'bar' }, measurements={ 'x': 42 })
tc.flush()

Configuring context for a telemetry client instance

.. code:: python

from applicationinsights import TelemetryClient
tc = TelemetryClient('<YOUR INSTRUMENTATION KEY GOES HERE>')
tc.context.application.ver = '1.2.3'
tc.context.device.id = 'My current device'
tc.context.device.oem_name = 'Asus'
tc.context.device.model = 'X31A'
tc.context.device.type = "Other"
tc.context.user.id = '[email protected]'
tc.context.properties['my_property'] = 'my_value'
tc.track_trace('My trace with context')
tc.flush()

Establishing correlation between telemetry objects

context field called operation_id can be set to associate telemetry items. Since operation_id is being set as a property of telemetry client, the client shouldn't be reused in parallel threads as it might lead to concurrency issues.

.. code:: python

tc = TelemetryClient(instrumentation_key=instrumentation_key)
tc.context.operation.id = <operation_id>
tc.track_trace('Test trace')
tc.flush()

Configuring channel related properties

.. code:: python

from applicationinsights import TelemetryClient
tc = TelemetryClient('<YOUR INSTRUMENTATION KEY GOES HERE>')
# flush telemetry every 30 seconds (assuming we don't hit max_queue_item_count first)
tc.channel.sender.send_interval_in_milliseconds = 30 * 1000
# flush telemetry if we have 10 or more telemetry items in our queue
tc.channel.queue.max_queue_length = 10

Configuring TelemetryProcessor

.. code:: python

from applicationinsights import TelemetryClient
def process(data, context):
   data.properties["NEW_PROP"] = "MYPROP"  # Add property
   context.user.id = "MYID"   # Change ID
   return True # Not filtered
tc = TelemetryClient('<YOUR INSTRUMENTATION KEY GOES HERE>')
tc.add_telemetry_processor(process)

Basic logging configuration (first option)

.. code:: python

import logging
from applicationinsights.logging import enable

# set up logging
enable('<YOUR INSTRUMENTATION KEY GOES HERE>')

# log something (this will be sent to the Application Insights service as a trace)
logging.info('This is a message')

# logging shutdown will cause a flush of all un-sent telemetry items
logging.shutdown()

Basic logging configuration (second option)

.. code:: python

import logging
from applicationinsights.logging import LoggingHandler

# set up logging
handler = LoggingHandler('<YOUR INSTRUMENTATION KEY GOES HERE>')
logging.basicConfig(handlers=[ handler ], format='%(levelname)s: %(message)s', level=logging.DEBUG)

# log something (this will be sent to the Application Insights service as a trace)
logging.debug('This is a message')

try:
    raise Exception('Some exception')
except:
    # this will send an exception to the Application Insights service
    logging.exception('Code went boom!')

# logging shutdown will cause a flush of all un-sent telemetry items
# alternatively flush manually via handler.flush()
logging.shutdown()

Advanced logging configuration

.. code:: python

import logging
from applicationinsights import channel
from applicationinsights.logging import LoggingHandler

# set up channel with context
telemetry_channel = channel.TelemetryChannel()
telemetry_channel.context.application.ver = '1.2.3'
telemetry_channel.context.properties['my_property'] = 'my_value'

# set up logging
handler = LoggingHandler('<YOUR INSTRUMENTATION KEY GOES HERE>', telemetry_channel=telemetry_channel)
handler.setLevel(logging.DEBUG)
handler.setFormatter(logging.Formatter('%(levelname)s: %(message)s'))
my_logger = logging.getLogger('simple_logger')
my_logger.setLevel(logging.DEBUG)
my_logger.addHandler(handler)

# log something (this will be sent to the Application Insights service as a trace)
my_logger.debug('This is a message')

# logging shutdown will cause a flush of all un-sent telemetry items
# alternatively flush manually via handler.flush()
logging.shutdown()

Logging unhandled exceptions

.. code:: python

from applicationinsights.exceptions import enable

# set up exception capture
enable('<YOUR INSTRUMENTATION KEY GOES HERE>')

# raise an exception (this will be sent to the Application Insights service as an exception telemetry object)
raise Exception('Boom!')

# exceptions will cause a flush of all un-sent telemetry items

Logging unhandled exceptions with context

.. code:: python

from applicationinsights import channel
from applicationinsights.exceptions import enable

# set up channel with context
telemetry_channel = channel.TelemetryChannel()
telemetry_channel.context.application.ver = '1.2.3'
telemetry_channel.context.properties['my_property'] = 'my_value'

# set up exception capture
enable('<YOUR INSTRUMENTATION KEY GOES HERE>', telemetry_channel=telemetry_channel)

# raise an exception (this will be sent to the Application Insights service as an exception telemetry object)
raise Exception('Boom!')

# exceptions will cause a flush of all un-sent telemetry items

Integrating with Flask

.. code:: python

from flask import Flask
from applicationinsights.flask.ext import AppInsights

# instantiate the Flask application
app = Flask(__name__)
app.config['APPINSIGHTS_INSTRUMENTATIONKEY'] = '<YOUR INSTRUMENTATION KEY GOES HERE>'

# log requests, traces and exceptions to the Application Insights service
appinsights = AppInsights(app)

# define a simple route
@app.route('/')
def hello_world():
    # the following message will be sent to the Flask log as well as Application Insights
    app.logger.info('Hello World route was called')

    return 'Hello World!'

# run the application
if __name__ == '__main__':
    app.run()

Integrating with Django

Place the following in your settings.py file:

.. code:: python

# If on Django < 1.10
MIDDLEWARE_CLASSES = [
    # ... or whatever is below for you ...
    'django.middleware.security.SecurityMiddleware',
    'django.contrib.sessions.middleware.SessionMiddleware',
    'django.middleware.common.CommonMiddleware',
    'django.middleware.csrf.CsrfViewMiddleware',
    'django.contrib.auth.middleware.AuthenticationMiddleware',
    'django.contrib.auth.middleware.SessionAuthenticationMiddleware',
    'django.contrib.messages.middleware.MessageMiddleware',
    'django.middleware.clickjacking.XFrameOptionsMiddleware',
    # ... or whatever is above for you ...
    'applicationinsights.django.ApplicationInsightsMiddleware',   # Add this middleware to the end
]

# If on Django >= 1.10
MIDDLEWARE = [
    # ... or whatever is below for you ...
    'django.middleware.security.SecurityMiddleware',
    'django.contrib.sessions.middleware.SessionMiddleware',
    'django.middleware.common.CommonMiddleware',
    'django.middleware.csrf.CsrfViewMiddleware',
    'django.contrib.auth.middleware.AuthenticationMiddleware',
    'django.contrib.messages.middleware.MessageMiddleware',
    'django.middleware.clickjacking.XFrameOptionsMiddleware',
    # ... or whatever is above for you ...
    'applicationinsights.django.ApplicationInsightsMiddleware',   # Add this middleware to the end
]

APPLICATION_INSIGHTS = {
    # (required) Your Application Insights instrumentation key
    'ikey': "00000000-0000-0000-0000-000000000000",
    
    # (optional) By default, request names are logged as the request method
    # and relative path of the URL.  To log the fully-qualified view names
    # instead, set this to True.  Defaults to False.
    'use_view_name': True,
    
    # (optional) To log arguments passed into the views as custom properties,
    # set this to True.  Defaults to False.
    'record_view_arguments': True,
    
    # (optional) Exceptions are logged by default, to disable, set this to False.
    'log_exceptions': False,
    
    # (optional) Events are submitted to Application Insights asynchronously.
    # send_interval specifies how often the queue is checked for items to submit.
    # send_time specifies how long the sender waits for new input before recycling
    # the background thread.
    'send_interval': 1.0, # Check every second
    'send_time': 3.0, # Wait up to 3 seconds for an event
    
    # (optional, uncommon) If you must send to an endpoint other than the
    # default endpoint, specify it here:
    'endpoint': "https://dc.services.visualstudio.com/v2/track",
}

This will log all requests and exceptions to the instrumentation key specified in the APPLICATION_INSIGHTS setting. In addition, an appinsights property will be placed on each incoming request object in your views. This will have the following properties:

• client: This is an instance of the applicationinsights.TelemetryClient type, which will submit telemetry to the same instrumentation key, and will parent each telemetry item to the current request.
• request: This is the applicationinsights.channel.contracts.RequestData instance for the current request. You can modify properties on this object during the handling of the current request. It will be submitted when the request has finished.
• context: This is the applicationinsights.channel.TelemetryContext object for the current ApplicationInsights sender.

You can also hook up logging to Django. For example, to log all builtin Django warnings and errors, use the following logging configuration in settings.py:

.. code:: python

LOGGING = {
    'version': 1,
    'disable_existing_loggers': False,
    'handlers': {
        # The application insights handler is here
        'appinsights': {
            'class': 'applicationinsights.django.LoggingHandler',
            'level': 'WARNING'
        }
    },
    'loggers': {
        'django': {
            'handlers': ['appinsights'],
            'level': 'WARNING',
            'propagate': True,
        }
    }
}

See Django's logging documentation <https://docs.djangoproject.com/en/1.11/topics/logging/>_ for more information.

Integrating with other web frameworks

For any other Python web framework that is WSGI compliant <https://www.python.org/dev/peps/pep-0333/>, the WSGIApplication <https://github.com/Microsoft/ApplicationInsights-Python/blob/master/applicationinsights/requests/WSGIApplication.py> can be used as a middleware to log requests to Application Insights.

Add common properties to WSGIApplication request events by passing in a dictionary to the WSGIApplication constructor:

.. code:: python

from wsgiref.simple_server import make_server
from pyramid.config import Configurator
from pyramid.response import Response
from applicationinsights.requests import WSGIApplication

# define a simple pyramid route
def hello_world(request):
    return Response('Hello World!')

# construct dictionary which contains properties to be included with every request event
common_properties = {
    "service": "hello_world_flask_app",
    "environment": "production"
}

if __name__ == '__main__':
    # create a simple pyramid app
    with Configurator() as config:
        config.add_route('hello', '/')
        config.add_view(hello_world, route_name='hello')
        app = config.make_wsgi_app()

        # wrap the app in the application insights request logging middleware
        app = WSGIApplication('<YOUR INSTRUMENTATION KEY GOES HERE>', app, common_properties=common_properties)

    # run the app
    server = make_server('0.0.0.0', 6543, app)
    server.serve_forever()