signalfx-collect

SignalFx library to collect nodejs metrics and report to SignalFx

Usage no npm install needed!

<script type="module">
  import signalfxCollect from 'https://cdn.skypack.dev/signalfx-collect';
</script>

README

GitHub release (latest by date) GitHub branch checks state npm node-current

SignalFx Node.js Collect

The SignalFx Node.js Collect is a library to monitor a Node.js application. It collects metrics about CPU utilization, memory usage, event loop, memory leaks, garbage collection and HTTP requests.

Installation

To install using npm:

$ npm install signalfx-collect

Note: You will get an error when > node-pre-gyp install --fallback-to-build is executed, stating that pre-built binaries couldn't be downloaded. We do not provide pre-built binaries. To ensure that no one can exploit binaries fetching we point node-pre-gyp to one of our domains, to a path which returns 403.

Usage

To use this library, you need a SignalFx org access token. For more information on access tokens, see the API's Authentication documentation. If your organization is currently sending metrics to SignalFx through SignalFx Node.js client, you can provide a preconfigured client object instead of an org access token.

The code example below is how to import signalfx-collect and start collecting metrics. Collect HTTP request metrics by registering middleware to your express application as shown in the example code below.

const express = require('express');
const app = express();

const SignalFxCollect = require('signalfx-collect');

let config = {
    accessToken: 'MY_ACCESS_TOKEN',
    interval: 1000
};

const collect = new SignalFxCollect(config);
collect.start();

// Add a middleware from SignalFxCollect module to collect HTTP metrics
app.use(collect.getMiddleware('express'));
app.get('/hello', (req, res) => {
    res.send('world');
});

To authenticate with a client object rather than an org access token, use the following example code:

...

const signalfx = require('signalfx');
const client = new signalfx.Ingest('MY_ACCESS_TOKEN');

let config = {
    signalFxClient: client,
    interval: 1000
};
const collect = new SignalFxCollect(config);
collect.start();

...

Once collect.start() is called, signalfx-collect will start polling for the metrics based on a provided interval in milliseconds.

When creating a neew SignalFxCollect instance, you must pass the config object which may contain the following fields:

  • accessToken (string) - SignalFx access token as explained above.
  • signalFxClient (string) - SignalFx Ingest client object.
  • ingestEndpoint (string) - custom url to send datapoints in format http://custom.domain/api/path
  • interval (int) - interval rate in milliseconds. (default: 10000)
  • extraDimensions (dict) - a dictionary of additional dimensions sent with metrics and events with string keys/values. Default is an empty dictionary.
  • logLevel (string) - a level for logging verbosity. Available levels are debug, info. No log will be shown except system exceptions if the level is not specified.

Below are the metrics sent by SignalFx Node.js Collect:

Metrics

  • nodejs.cpu.utilization.system (gauge)
  • nodejs.cpu.utilization.user (gauge)
  • nodejs.cpu.utilization.total (gauge)
  • nodejs.memory.system.total (gauge)
  • nodejs.memory.system.free (gauge)
  • nodejs.memory.heap.total (gauge)
  • nodejs.memory.heap.used (gauge)
  • nodejs.memory.rss (gauge)
  • nodejs.memory.gc.size (cumulative_counter)
  • nodejs.memory.gc.pause (cumulative_counter)
  • nodejs.memory.gc.total (cumulative_counter)
  • nodejs.memory.heap.leak (cumulative_counter)
  • nodejs.event_loop.max (gauge)
  • nodejs.event_loop.min (gauge)
  • nodejs.http.rq_total (cumulative_counter)
  • nodejs.http.rq_time (gauge)
  • nodejs.http.rq_<status_code> (cumulative_counter)
  • nodejs.http.rq_size (gauge)

License

Apache Software License v2 © SignalFx