UNB/ CS/ David Bremner/ teaching/ cs2613/ books/ nodejs-api/ perf hooks

Performance measurement APIs

Stability: 2 - Stable

This module provides an implementation of a subset of the W3C Web Performance APIs as well as additional APIs for Node.js-specific performance measurements.

Node.js supports the following Web Performance APIs:

const { PerformanceObserver, performance } = require('perf_hooks');

const obs = new PerformanceObserver((items) => {
  console.log(items.getEntries()[0].duration);
  performance.clearMarks();
});
obs.observe({ entryTypes: ['measure'] });
performance.measure('Start to Now');

performance.mark('A');
doSomeLongRunningProcess(() => {
  performance.measure('A to Now', 'A');

  performance.mark('B');
  performance.measure('A to B', 'A', 'B');
});

perf_hooks.performance

An object that can be used to collect performance metrics from the current Node.js instance. It is similar to window.performance in browsers.

performance.clearMarks([name])

If name is not provided, removes all PerformanceMark objects from the Performance Timeline. If name is provided, removes only the named mark.

performance.eventLoopUtilization([utilization1[, utilization2]])

The eventLoopUtilization() method returns an object that contains the cumulative duration of time the event loop has been both idle and active as a high resolution milliseconds timer. The utilization value is the calculated Event Loop Utilization (ELU).

If bootstrapping has not yet finished on the main thread the properties have the value of 0. The ELU is immediately available on Worker threads since bootstrap happens within the event loop.

Both utilization1 and utilization2 are optional parameters.

If utilization1 is passed, then the delta between the current call's active and idle times, as well as the corresponding utilization value are calculated and returned (similar to process.hrtime()).

If utilization1 and utilization2 are both passed, then the delta is calculated between the two arguments. This is a convenience option because, unlike process.hrtime(), calculating the ELU is more complex than a single subtraction.

ELU is similar to CPU utilization, except that it only measures event loop statistics and not CPU usage. It represents the percentage of time the event loop has spent outside the event loop's event provider (e.g. epoll_wait). No other CPU idle time is taken into consideration. The following is an example of how a mostly idle process will have a high ELU.

'use strict';
const { eventLoopUtilization } = require('perf_hooks').performance;
const { spawnSync } = require('child_process');

setImmediate(() => {
  const elu = eventLoopUtilization();
  spawnSync('sleep', ['5']);
  console.log(eventLoopUtilization(elu).utilization);
});

Although the CPU is mostly idle while running this script, the value of utilization is 1. This is because the call to child_process.spawnSync() blocks the event loop from proceeding.

Passing in a user-defined object instead of the result of a previous call to eventLoopUtilization() will lead to undefined behavior. The return values are not guaranteed to reflect any correct state of the event loop.

performance.mark([name])

Creates a new PerformanceMark entry in the Performance Timeline. A PerformanceMark is a subclass of PerformanceEntry whose performanceEntry.entryType is always 'mark', and whose performanceEntry.duration is always 0. Performance marks are used to mark specific significant moments in the Performance Timeline.

performance.measure(name[, startMark[, endMark]])

Creates a new PerformanceMeasure entry in the Performance Timeline. A PerformanceMeasure is a subclass of PerformanceEntry whose performanceEntry.entryType is always 'measure', and whose performanceEntry.duration measures the number of milliseconds elapsed since startMark and endMark.

The startMark argument may identify any existing PerformanceMark in the Performance Timeline, or may identify any of the timestamp properties provided by the PerformanceNodeTiming class. If the named startMark does not exist, then startMark is set to timeOrigin by default.

The optional endMark argument must identify any existing PerformanceMark in the Performance Timeline or any of the timestamp properties provided by the PerformanceNodeTiming class. endMark will be performance.now() if no parameter is passed, otherwise if the named endMark does not exist, an error will be thrown.

performance.nodeTiming

This property is an extension by Node.js. It is not available in Web browsers.

An instance of the PerformanceNodeTiming class that provides performance metrics for specific Node.js operational milestones.

performance.now()

Returns the current high resolution millisecond timestamp, where 0 represents the start of the current node process.

performance.timeOrigin

The timeOrigin specifies the high resolution millisecond timestamp at which the current node process began, measured in Unix time.

performance.timerify(fn)

This property is an extension by Node.js. It is not available in Web browsers.

Wraps a function within a new function that measures the running time of the wrapped function. A PerformanceObserver must be subscribed to the 'function' event type in order for the timing details to be accessed.

const {
  performance,
  PerformanceObserver
} = require('perf_hooks');

function someFunction() {
  console.log('hello world');
}

const wrapped = performance.timerify(someFunction);

const obs = new PerformanceObserver((list) => {
  console.log(list.getEntries()[0].duration);
  obs.disconnect();
});
obs.observe({ entryTypes: ['function'] });

// A performance timeline entry will be created
wrapped();

performance.eventLoopUtilization([util1][,util2])

The eventLoopUtilization() method returns an object that contains the cumulative duration of time the event loop has been both idle and active as a high resolution milliseconds timer. The utilization value is the calculated Event Loop Utilization (ELU). If bootstrapping has not yet finished, the properties have the value of 0.

util1 and util2 are optional parameters.

If util1 is passed then the delta between the current call's active and idle times are calculated and returned (similar to process.hrtime()). Likewise the adjusted utilization value is calculated.

If util1 and util2 are both passed then the calculation adjustments are done between the two arguments. This is a convenience option because unlike process.hrtime() additional work is done to calculate the ELU.

ELU is similar to CPU utilization except that it is calculated using high precision wall-clock time. It represents the percentage of time the event loop has spent outside the event loop's event provider (e.g. epoll_wait). No other CPU idle time is taken into consideration. The following is an example of how a mostly idle process will have a high ELU.

'use strict';
const { eventLoopUtilization } = require('perf_hooks').performance;
const { spawnSync } = require('child_process');

setImmediate(() => {
  const elu = eventLoopUtilization();
  spawnSync('sleep', ['5']);
  console.log(eventLoopUtilization(elu).utilization);
});

Although the CPU is mostly idle while running this script, the value of utilization is 1. This is because the call to child_process.spawnSync() blocks the event loop from proceeding.

Passing in a user-defined object instead of the result of a previous call to eventLoopUtilization() will lead to undefined behavior. The return values are not guaranteed to reflect any correct state of the event loop.

Class: PerformanceEntry

performanceEntry.duration

The total number of milliseconds elapsed for this entry. This value will not be meaningful for all Performance Entry types.

performanceEntry.name

The name of the performance entry.

performanceEntry.startTime

The high resolution millisecond timestamp marking the starting time of the Performance Entry.

performanceEntry.entryType

The type of the performance entry. It may be one of:

performanceEntry.kind

This property is an extension by Node.js. It is not available in Web browsers.

When performanceEntry.entryType is equal to 'gc', the performance.kind property identifies the type of garbage collection operation that occurred. The value may be one of:

performanceEntry.flags

This property is an extension by Node.js. It is not available in Web browsers.

When performanceEntry.entryType is equal to 'gc', the performance.flags property contains additional information about garbage collection operation. The value may be one of:

Class: PerformanceNodeTiming extends PerformanceEntry

This property is an extension by Node.js. It is not available in Web browsers.

Provides timing details for Node.js itself. The constructor of this class is not exposed to users.

performanceNodeTiming.bootstrapComplete

The high resolution millisecond timestamp at which the Node.js process completed bootstrapping. If bootstrapping has not yet finished, the property has the value of -1.

performanceNodeTiming.environment

The high resolution millisecond timestamp at which the Node.js environment was initialized.

performanceNodeTiming.loopExit

The high resolution millisecond timestamp at which the Node.js event loop exited. If the event loop has not yet exited, the property has the value of -1. It can only have a value of not -1 in a handler of the 'exit' event.

performanceNodeTiming.loopStart

The high resolution millisecond timestamp at which the Node.js event loop started. If the event loop has not yet started (e.g., in the first tick of the main script), the property has the value of -1.

performanceNodeTiming.nodeStart

The high resolution millisecond timestamp at which the Node.js process was initialized.

performanceNodeTiming.v8Start

The high resolution millisecond timestamp at which the V8 platform was initialized.

performanceNodeTiming.idleTime

The high resolution millisecond timestamp of the amount of time the event loop has been idle within the event loop's event provider (e.g. epoll_wait). This does not take CPU usage into consideration. If the event loop has not yet started (e.g., in the first tick of the main script), the property has the value of 0.

Class: perf_hooks.PerformanceObserver

new PerformanceObserver(callback)

PerformanceObserver objects provide notifications when new PerformanceEntry instances have been added to the Performance Timeline.

const {
  performance,
  PerformanceObserver
} = require('perf_hooks');

const obs = new PerformanceObserver((list, observer) => {
  console.log(list.getEntries());
  observer.disconnect();
});
obs.observe({ entryTypes: ['mark'], buffered: true });

performance.mark('test');

Because PerformanceObserver instances introduce their own additional performance overhead, instances should not be left subscribed to notifications indefinitely. Users should disconnect observers as soon as they are no longer needed.

The callback is invoked when a PerformanceObserver is notified about new PerformanceEntry instances. The callback receives a PerformanceObserverEntryList instance and a reference to the PerformanceObserver.

performanceObserver.disconnect()

Disconnects the PerformanceObserver instance from all notifications.

performanceObserver.observe(options)

Subscribes the PerformanceObserver instance to notifications of new PerformanceEntry instances identified by options.entryTypes.

When options.buffered is false, the callback will be invoked once for every PerformanceEntry instance:

const {
  performance,
  PerformanceObserver
} = require('perf_hooks');

const obs = new PerformanceObserver((list, observer) => {
  // Called three times synchronously. `list` contains one item.
});
obs.observe({ entryTypes: ['mark'] });

for (let n = 0; n < 3; n++)
  performance.mark(`test${n}`);
const {
  performance,
  PerformanceObserver
} = require('perf_hooks');

const obs = new PerformanceObserver((list, observer) => {
  // Called once. `list` contains three items.
});
obs.observe({ entryTypes: ['mark'], buffered: true });

for (let n = 0; n < 3; n++)
  performance.mark(`test${n}`);

Class: PerformanceObserverEntryList

The PerformanceObserverEntryList class is used to provide access to the PerformanceEntry instances passed to a PerformanceObserver. The constructor of this class is not exposed to users.

performanceObserverEntryList.getEntries()

Returns a list of PerformanceEntry objects in chronological order with respect to performanceEntry.startTime.

performanceObserverEntryList.getEntriesByName(name[, type])

Returns a list of PerformanceEntry objects in chronological order with respect to performanceEntry.startTime whose performanceEntry.name is equal to name, and optionally, whose performanceEntry.entryType is equal to type.

performanceObserverEntryList.getEntriesByType(type)

Returns a list of PerformanceEntry objects in chronological order with respect to performanceEntry.startTime whose performanceEntry.entryType is equal to type.

perf_hooks.monitorEventLoopDelay([options])

This property is an extension by Node.js. It is not available in Web browsers.

Creates a Histogram object that samples and reports the event loop delay over time. The delays will be reported in nanoseconds.

Using a timer to detect approximate event loop delay works because the execution of timers is tied specifically to the lifecycle of the libuv event loop. That is, a delay in the loop will cause a delay in the execution of the timer, and those delays are specifically what this API is intended to detect.

const { monitorEventLoopDelay } = require('perf_hooks');
const h = monitorEventLoopDelay({ resolution: 20 });
h.enable();
// Do something.
h.disable();
console.log(h.min);
console.log(h.max);
console.log(h.mean);
console.log(h.stddev);
console.log(h.percentiles);
console.log(h.percentile(50));
console.log(h.percentile(99));

Class: Histogram

Tracks the event loop delay at a given sampling rate. The constructor of this class not exposed to users.

This property is an extension by Node.js. It is not available in Web browsers.

histogram.disable()

Disables the event loop delay sample timer. Returns true if the timer was stopped, false if it was already stopped.

histogram.enable()

Enables the event loop delay sample timer. Returns true if the timer was started, false if it was already started.

histogram.exceeds

The number of times the event loop delay exceeded the maximum 1 hour event loop delay threshold.

histogram.max

The maximum recorded event loop delay.

histogram.mean

The mean of the recorded event loop delays.

histogram.min

The minimum recorded event loop delay.

histogram.percentile(percentile)

Returns the value at the given percentile.

histogram.percentiles

Returns a Map object detailing the accumulated percentile distribution.

histogram.reset()

Resets the collected histogram data.

histogram.stddev

The standard deviation of the recorded event loop delays.

Examples

Measuring the duration of async operations

The following example uses the Async Hooks and Performance APIs to measure the actual duration of a Timeout operation (including the amount of time it took to execute the callback).

'use strict';
const async_hooks = require('async_hooks');
const {
  performance,
  PerformanceObserver
} = require('perf_hooks');

const set = new Set();
const hook = async_hooks.createHook({
  init(id, type) {
    if (type === 'Timeout') {
      performance.mark(`Timeout-${id}-Init`);
      set.add(id);
    }
  },
  destroy(id) {
    if (set.has(id)) {
      set.delete(id);
      performance.mark(`Timeout-${id}-Destroy`);
      performance.measure(`Timeout-${id}`,
                          `Timeout-${id}-Init`,
                          `Timeout-${id}-Destroy`);
    }
  }
});
hook.enable();

const obs = new PerformanceObserver((list, observer) => {
  console.log(list.getEntries()[0]);
  performance.clearMarks();
  observer.disconnect();
});
obs.observe({ entryTypes: ['measure'], buffered: true });

setTimeout(() => {}, 1000);

Measuring how long it takes to load dependencies

The following example measures the duration of require() operations to load dependencies:

'use strict';
const {
  performance,
  PerformanceObserver
} = require('perf_hooks');
const mod = require('module');

// Monkey patch the require function
mod.Module.prototype.require =
  performance.timerify(mod.Module.prototype.require);
require = performance.timerify(require);

// Activate the observer
const obs = new PerformanceObserver((list) => {
  const entries = list.getEntries();
  entries.forEach((entry) => {
    console.log(`require('${entry[0]}')`, entry.duration);
  });
  obs.disconnect();
});
obs.observe({ entryTypes: ['function'], buffered: true });

require('some-module');