lambda-powertools-http-client

HTTP client that automatically forwards correlation IDs (captured via @michaelfecher/lambda-powertools-correlation-ids), and follows DAZN's convention around recording metrics around integration points.

Main features:

• auto-forwards any correlation IDs captured with the @michaelfecher/lambda-powertools-correlation-ids package as HTTP headers

• auto-record custom metrics using the @michaelfecher/datadog-metrics package, which defaults to async mode (i.e. writing to stdout in DogStatsD format) but can be configured via the DATADOG_METRICS_MODE environment variable

• custom metrics include:

  • {hostName}.response.latency [histogram]: e.g. google.com.response.latency

  • {hostName}.response.{statusCode} [count]: e.g. google.com.response.200

  metric names can be overriden with the metricName option (see below for details)

• all custom metrics include the tags awsRegion, functionName, functionVersion, method (e.g. POST) and path (e.g. /v1/signin)

• you can add additional tags by passing them in via the metricTags option (see below for details)

• supports timeout

Getting Started

Install from NPM: npm install @michaelfecher/lambda-powertools-http-client

API

Basic usage looks like this:

const HTTP = require('@michaelfecher/lambda-powertools-http-client')

const sayIt = async () => {
  const httpRequest = {
    uri: `https://example.com/dev/say`,
    method: 'post',
    body: { message: 'hello world' }
  }

  await HTTP(httpRequest)
}

It's essentially a function that accepts a request of type:

{
  uri/url : string (either uri or url must be specified)
  method  : GET (default) | POST | PUT | HEAD | DELETE | PATCH
  headers : object
  qs      : object
  body    : object
  metricName [optional] : string     // override the default metric name, e.g. 'adyenApi', which changes metrics to 'adyenapi.latency' and 'adyenapi.202'
  metricTags [optional] : string []  // additional tags for metrics, e.g. ['request_type:submit', 'load_test']
  timeout [optional] : int (millis)
}