The core measured library that has the Metric interfaces and implementations.
npm install measured-core
The core library has the following metrics classes:
Values that can be read instantly via a supplied call back.
Just like a Gauge but its value is set directly rather than supplied by a callback.
Like a mix of the regular and settable Gauge it takes a call back that returns a promise that will resolve the cached value and an interval that it should call the callback on to update its cached value.
Counters are things that increment or decrement.
Timers are a combination of Meters and Histograms. They measure the rate as well as distribution of scalar events.
Keeps a reservoir of statistically relevant values to explore their distribution.
Things that are measured as events / interval.
The core library comes with a basic registry class
that is not aware of dimensions / tags and leaves reporting up to you.
See the measured-reporting module for more advanced and featured registries.
See The measured-core modules for the full list of exports for require('measured-core').
Step 1: Add measurements to your code. For example, lets track the requests/sec of a http server:
var http = require('http');
var stats = require('measured').createCollection();
http.createServer(function(req, res) {
stats.meter('requestsPerSecond').mark();
res.end('Thanks');
}).listen(3000);Step 2: Show the collected measurements (more advanced examples follow later):
setInterval(function() {
console.log(stats.toJSON());
}, 1000);This will output something like this every second:
{ requestsPerSecond:
{ mean: 1710.2180279856818,
count: 10511,
'currentRate': 1941.4893498239829,
'1MinuteRate': 168.08263156623656,
'5MinuteRate': 34.74630977619571,
'15MinuteRate': 11.646507524106095 } }
Step 3: Aggregate the data into your backend of choice. Here are a few time series data aggregators.