Skip to content

Node.js bindings to Librato that provide advanced statistics

License

Notifications You must be signed in to change notification settings

eremzeit/nodebrato

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

20 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Nodebrato

A node.js bindings for Librato metrics that provides advanced statistics which allow you to reduce your reporting frequency and ultimately lower your montly Librato bill.

It was originally created for http://showgoers.tv but I forked it out of that codebase to open it up for community use and contributions.

Features

  • Supports pre-registering metric definitions
  • Can update metric definitions automatically so that you don't have to do that manually in the librato interface. (edit: this feature hasn't quite been merged into master. Pull-requests are welcome!)
  • Gives more control over how each individual metric is collected, aggregated and submitted
    • Reporting intervals can be defined on a per-metric basis. This can can save you money by allowing you to only report when you need to.
    • Supports defining separate client-side aggregation functions and librato-side aggregation functions
      • This lets you use aggregation functions that librato doesn't support (eg. advanced stastistics and quantiles)
      • By giving you more additional descriptive statistics, you can drastically increase your reporting period (less $$$) but still have a clear understanding of that stat.
      • It's also useful for librato power users who make heavy use of alerts and composite functions.
  • Supports Librato graph annotations (eg. for marking deployments, etc)

How is this different from librato-node?

  • librato-node aggregates all measurements inline, which limits flexibility but is more suited for extremely high performance reporting.
  • librato-node is written in Coffeescript.

Example

let metricDefinitions = {
  'errors': {
    libratoAggFunction: 'sum',
    periodMs: 10000 //perhaps we want errors reported at a higher resolution than other metrics
  },

  'star_rating': {
    libratoAggFunction: 'average',
  },

  'web_requests': {
    clientAggFunction: 'sum',
    libratoAggFunction: 'average',
    libratoMetricProperties: {
      display_name: 'Site Requests',
      description: 'The number of requests made to the web server',
      attributes: {
        color: '#ff0000'
      }
    },
  },

  'response_time_ms': {
    clientAggFunction: 'quantiles',
    libratoAggFunction: 'min',
    quantiles: [0, .1, .90, 1], //when submitted to librato, will actually create 4 separate metrics (ie. response_time_ms.q0, response_time_ms.q10, response_time_ms.q90, response_time_ms.q100)
    periodMs: 30 * 60 * 1000  //because we're intelligently aggregating, we only need to report every thirty minutes
  }
}

let librato = new Librato({
  source: 'my_default_source',
  definitions: metricDefinitions,
  logging: true, //turn on debug output to console
  periodMs: 10000
})

librato.start()

librato.increment('requests', 10)
librato.measure('response_time_ms', 1)
librato.measure('star_rating', 5)
librato.measure('star_rating', 4, 'another_source')

librato.measure('not_defined_metric', 1)  //when using the measure method will default to "mean" as an aggregation function
librato.increment('not_defined_count', 1)  //when using the increment method will default to "sum" as an aggregation function

About

Node.js bindings to Librato that provide advanced statistics

Resources

License

Stars

Watchers

Forks

Packages

No packages published