A simple node.js module to count, report, and plot application metrics.
npm install abacus
var Abacus = require('abacus');
var config = {
statsD: {
connection:{
host: 'localhost',
port: 5007
},
// This is the prefix for all metric names sent to statsD / graphite
metricPrefix: 'apps.abacus.'
},
flushFrequency: 60000,
printFrequency: 10000,
resetOnFlush: true, // set all values to 0 on flush
deleteOnFlush: false, // delete all values on flush (will not send 0's)
printEmpty: true, // whether or not to print empty objects (default false)
printOnFlush: false, // print the metrics just before flushing them (default false)
debug: console.log // Prints messages from abacus to STDOUT
};
var metrics = new Abacus(config);
// increment by 1
metrics.increment('metricName');
// increment by 5
metrics.increment('metricName', 5);
// set value of metric
metrics.set('metricName', 153);
// Get value of metric
metrics.get('metricName');
// Print a summary of counters to STDOUT every 10 seconds. This is automatically setup if config.printFrequency is set.
metrics.printPeriodically(10000);
// Periodically flush counters to statsD. This is opposed to sending the counter each time it is changed
// If config.resetOnFlush is false, the counters will be cumulative. Otherwise, they're reset to 0 on each flush.
// This is automatically setup if config.flushFrequency is set.
metrics.flushPeriodically(60000); // in milliseconds. Flush every 60 seconds
You can pass a configuration object directly to abacus on instantiation: new Abacus(config);
. See usage example.
Abacus can be configured to send its counters to an instance of StatsD. You can configure the StatsD connection as shown in the example.
The debug
option allows you to pass in a function that will be invoked each time abacus logs something. If not set, the log will be silently discarded.