Skip to content

Latest commit

 

History

History
130 lines (91 loc) · 5.42 KB

_index.md

File metadata and controls

130 lines (91 loc) · 5.42 KB
title
Install Serverless Monitoring for Azure Functions

Overview

This page explains how to collect traces, trace metrics, runtime metrics, and custom metrics from your Azure Functions. To collect additional metrics, install the Datadog Azure integration.

Setup

{{< programming-lang-wrapper langs="nodejs,python" >}} {{< programming-lang lang="nodejs" >}}

  1. Install dependencies. Run the following commands:

    npm install @datadog/serverless-compat
npm install dd-trace

    To use automatic instrumentation, you must use dd-trace v5.25+.

    Datadog recommends pinning the package versions and regularly upgrading to the latest versions of both @datadog/serverless-compat and dd-trace to ensure you have access to enhancements and bug fixes.

  2. Start the Datadog serverless compatibility layer and initialize the Node.js tracer. Add the following lines to your main application entry point file (for example, app.js):

    require('@datadog/serverless-compat').start();

// This line must come before importing any instrumented module. 
const tracer = require('dd-trace').init()

  3. (Optional) Enable runtime metrics. See Node.js Runtime Metrics.

  4. (Optional) Enable custom metrics. See Metric Submission: DogStatsD.

{{< /programming-lang >}} {{< programming-lang lang="python" >}}

  1. Install dependencies. Run the following commands:

    pip install datadog-serverless-compat
pip install ddtrace

    To use automatic instrumentation, you must use dd-trace v2.19+.

    Datadog recommends using the latest versions of both datadog-serverless-compat and ddtrace to ensure you have access to enhancements and bug fixes.

  2. Initialize the Datadog Python tracer and serverless compatibility layer. Add the following lines to your main application entry point file:

    from datadog_serverless_compat import start
from ddtrace import tracer, patch_all

start()
patch_all()

  3. (Optional) Enable runtime metrics. See Python Runtime Metrics.

  4. (Optional) Enable custom metrics. See Metric Submission: DogStatsD. 1: /tracing/trace_collection/automatic_instrumentation/?tab=singlestepinstrumentation 2: /tracing/metrics/runtime_metrics/python/ 3: /metrics/custom_metrics/dogstatsd_metrics_submission/?code-lang=python {{< /programming-lang >}} {{< /programming-lang-wrapper >}}

  5. Deploy your function.

  6. Configure Datadog intake. Add the following environment variables to your function's application settings:

    DD_API_KEY=<YOUR_API_KEY>
DD_SITE={{< region-param key=dd_site code="true" >}}
    • Set DD_API_KEY to your Datadog API key.
    • Set DD_SITE to {{< region-param key=dd_site code="true" >}}. (Ensure that your correct Datadog site is selected in the drop-down on the right side of this page.)

  7. Configure Unified Service Tagging. You can collect metrics from your Azure Functions by installing the Datadog Azure integration. To correlate these metrics with your traces, first set the env, service, and version tags on your resource in Azure. Then, configure the following environment variables. You can add custom tags as DD_TAGS.

    DD_ENV="<ENVIRONMENT>"
DD_SERVICE="<SERVICE_NAME>"
DD_VERSION="<VERSION>"
DD_TAGS="<KEY_1:VALUE_1>,<KEY_2:VALUE_2>"

What's next?

Enable/disable trace metrics

Trace metrics are enabled by default. To configure trace metrics, use the following environment variable:

DD_TRACE_STATS_COMPUTATION_ENABLED : Enables (true) or disables (false) trace metrics. Defaults to true.

Values: true, false

Troubleshooting

Enable debug logs

You can collect debug logs for troubleshooting. To configure debug logs, use the following environment variables:

DD_TRACE_DEBUG : Enables (true) or disables (false) debug logging for the Datadog Tracing Library. Defaults to false.

Values: true, false

DD_LOG_LEVEL : Sets logging level for the Datadog Serverless Compatibility Layer. Defaults to info.

Values: trace, debug, info, warn, error, critical, off

Linux Consumption plans and GitHub Actions

To use a GitHub Action to deploy to a Linux Consumption function, you must configure your workflow to use an Azure Service Principal for RBAC. See Using Azure Service Principal for RBAC as Deployment Credential.