Node.js library for Stackdriver Profiler. Continuous CPU and heap profiling to improve performance and reduce costs.
Clone or download
Latest commit ad5efa8 Nov 20, 2018

README.md

Google Cloud Profiler

NPM Version Build Status Dependency Status devDependency Status Known Vulnerabilities

Alpha. This is an Alpha release of Stackdriver Profiler Node.js profiling agent. This feature might be changed in backward-incompatible ways and is not recommended for production use. It is not subject to any SLA or deprecation policy.

Prerequisites

  1. Your application will need to be using Node.js version 6.12.3 or greater, Node.js 8.9.4 or greater, or Node.js 10.4.1 or greater. The profiler will not be enabled when using earlier versions of Node 6, 8, and 10 because the profiler is not stable with those versions of Node.js.

    • Versions of Node.js 6 prior to 6.12.3 are impacted by this issue, which can cause segmentation faults when heap profiling is enabled.
    • Versions of Node.js before Node.js 8 prior to 8.9.4 are impacted by this issue, which causes a memory leak when time profiling is enabled.
    • Versions of Node.js 10 prior to 10.4.1 are impacted by this issue, which can cause garbage collection to take several minutes when heap profiling is enabled.
  2. You will need a project in the Google Developers Console. Your application can run anywhere, but the profiler data is associated with a particular project.

  3. You will need to enable the Stackdriver Profiler API for your project.

Basic Set-up

  1. Install @google-cloud/profiler with npm or add to your package.json.

    # Install through npm while saving to the local 'package.json'
    npm install --save @google-cloud/profiler
  2. Include and start the profiler at the beginning of your application:

    var profiler = require('@google-cloud/profiler').start();

    Some environments require a configuration to be passed to the start() function. For more details on this, see instructions for running outside of Google Cloud Platform, on App Engine flexible environment, on Google Compute Engine, and on Google Container Engine.

  3. If you are running your application locally, or on a machine where you are using the Google Cloud SDK, make sure to log in with the application default credentials:

    gcloud beta auth application-default login

    Alternatively, you can set GOOGLE_APPLICATION_CREDENTIALS. For more details on this, see Running elsewhere

Configuration

See the default configuration for a list of possible configuration options. These options can be passed to the agent through the object argument to the start command shown below:

require('@google-cloud/profiler').start({disableTime: true});

Alternatively, you can provide the configuration through a config file. This can be useful if you want to load our module using --require on the command line (which requires and starts the agent) instead of editing your main script. The GCLOUD_PROFILER_CONFIG environment variable should point to your configuration file.

export GCLOUD_PROFILER_CONFIG=./path/to/your/profiler/configuration.js

Changing log level

The profiler writes log statements to the console log for diagnostic purposes. By default, the log level is set to warn. You can adjust this by setting logLevel in the config. Setting logLevel to 0 will disable logging, 1 sets log level to error, 2 sets it to warn, 3 sets it to info, and 4 sets it to debug.

So, for example, to start the profiler with the log level at debug, you would do this:

require('@google-cloud/profiler').start({logLevel: 4});

Disabling heap or time profile collection

By default, the profiler collects both heap profiles, which show memory allocations, and time profiles, which capture how much wall-clock time is spent in different locations of the code. Using the configuration, it is possible to disable the collection of either type of profile.

To disable time profile collection, set disableTime to true:

require('@google-cloud/profiler').start({disableTime: true});

To disable heap profile collection, set disableHeap to true:

require('@google-cloud/profiler').start({disableHeap: true});

Running on Google Cloud Platform

There are three different services that can host Node.js applications within Google Cloud Platform: Google App Engine flexible environment, Google Compute Engine, and Google Container Engine. After installing @google-cloud/profiler in your project and ensuring that the environment you are using uses Node.js version 6.12.3 or greater, or Node.js 8.9.4 or greater, follow the service-specific instructions to enable the profiler.

Running on App Engine flexible environment

To enable the profiling agent for a Node.js program running in the App Engine flexible environment, import the agent at the top of your application’s main script or entry point by including the following code snippet:

var profiler = require('@google-cloud/profiler').start();

You can specify which version of Node.js you're using by adding a snippet like the following to your package.json:

  "engines": {
    "node": ">=8.9.4"
  }

The above snippet will ensure that you're using 8.9.4 or greater.

Deploy your application to App Engine Flexible environment as usual.

Running on Google Compute Engine

To enable the profiling agent for a Node.js program running in the Google Compute Engine environment, import the agent at the top of your application’s main script or entry point by including the following code snippet:

require('@google-cloud/profiler').start({
  serviceContext: {
    service: 'your-service',
    version: '1.0.0'
  }
});

You may also need to download build-essential. An example of this would be:

apt-get update
apt-get install build-essential

Running on Google Container Engine

To enable the profiling agent for a Node.js program running in the Google Container Engine environment, import the agent at the top of your application’s main script or entry point by including the following code snippet:

require('@google-cloud/profiler').start({
  serviceContext: {
    service: 'your-service',
    version: '1.0.0'
  }
});

You may also need to add build-essential to your environment.

Running elsewhere

You can still use @google-cloud/profiler if your application is running outside of Google Cloud Platform, for example, running locally, on-premise, or on another cloud provider.

  1. You will need to specify your project id and the service you want the collected profiles to be associated with, and (optionally) the version of the service when starting the profiler:
  require('@google-cloud/profiler').start({
    projectId: 'project-id',
    serviceContext: {
      service: 'your-service',
      version: '1.0.0'
    }
  });
  1. You will need to provide credential for your application.

    • If you are running your application on a development machine or test environment where you are using the gcloud command line tools, and are logged using gcloud beta auth application-default login, you already have sufficient credentials, and a service account key is not required.

    • You can provide credentials via Application Default Credentials. This is the reccomended method. 1. Create a new JSON service account key. 2. Copy the key somewhere your application can access it. Be sure not to expose the key publicly. 3. Set the environment variable GOOGLE_APPLICATION_CREDENTIALS to the full path to the key. The profiler will automatically look for this environment variable.

    • You may set the keyFilename or credentials configuration field to the full path or contents to the key file, respectively. Setting either of these fields will override either setting GOOGLE_APPLICATION_CREDENTIALS or logging in using gcloud.

      This is how you would set keyFilename:

      require('@google-cloud/profiler').start({
        projectId: 'project-id',
        serviceContext: {
          service: 'your-service',
          version: '1.0.0'
        },
        keyFilename: '/path/to/keyfile'
      });

      This is how you would set credentials:

      require('@google-cloud/profiler').start({
        projectId: 'project-id',
        serviceContext: {
          service: 'your-service',
          version: '1.0.0'
        },
        credentials: {
          client_email: 'email',
          private_key: 'private_key'
        }
      });