Performance measurement APIs
History
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:
import { performance, PerformanceObserver } from 'node:perf_hooks'; const obs = new PerformanceObserver((items) => { console.log(items.getEntries()[0].duration); performance.clearMarks(); }); obs.observe({ type: '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'); });
const { PerformanceObserver, performance } = require('node:perf_hooks'); const obs = new PerformanceObserver((items) => { console.log(items.getEntries()[0].duration); }); obs.observe({ type: 'measure' }); performance.measure('Start to Now'); performance.mark('A'); (async function doSomeLongRunningProcess() { await new Promise((r) => setTimeout(r, 5000)); performance.measure('A to Now', 'A'); performance.mark('B'); performance.measure('A to B', 'A', 'B'); })();
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
History
performance object as the receiver.performance.clearMarks(name?): void
stringIf name is not provided, removes all PerformanceMark objects from the
Performance Timeline. If name is provided, removes only the named mark.
performance.clearMeasures
History
performance object as the receiver.performance.clearMeasures(name?): void
stringIf name is not provided, removes all PerformanceMeasure objects from the
Performance Timeline. If name is provided, removes only the named measure.
performance.clearResourceTimings
History
performance object as the receiver.performance.clearResourceTimings(name?): void
stringIf name is not provided, removes all PerformanceResourceTiming objects from
the Resource Timeline. If name is provided, removes only the named resource.
performance.eventLoopUtilization
History
perf_hooks.eventLoopUtilization alias.performance.eventLoopUtilization(utilization1?, utilization2?): Object
This is an alias of perf_hooks.eventLoopUtilization().
This property is an extension by Node.js. It is not available in Web browsers.
performance.getEntries
History
performance object as the receiver.performance.getEntries(): PerformanceEntry[]
PerformanceEntry[]Returns a list of PerformanceEntry objects in chronological order with
respect to performanceEntry.startTime. If you are only interested in
performance entries of certain types or that have certain names, see
performance.getEntriesByType() and performance.getEntriesByName().
performance.getEntriesByName
History
performance object as the receiver.performance.getEntriesByName(name, type?): PerformanceEntry[]
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.
performance.getEntriesByType
History
performance object as the receiver.performance.getEntriesByType(type): PerformanceEntry[]
stringPerformanceEntry[]Returns a list of PerformanceEntry objects in chronological order
with respect to performanceEntry.startTime whose performanceEntry.entryType
is equal to type.
performance.mark(name, options?): void
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.
The created PerformanceMark entry is put in the global Performance Timeline
and can be queried with performance.getEntries,
performance.getEntriesByName, and performance.getEntriesByType. When the
observation is performed, the entries should be cleared from the global
Performance Timeline manually with performance.clearMarks.
performance.markResourceTiming
History
performance.markResourceTiming(timingInfo, requestedUrl, initiatorType, global, cacheMode, bodyInfo, responseStatus, deliveryType?): void
ObjectstringstringObjectstringObjectnumberstring''.This property is an extension by Node.js. It is not available in Web browsers.
Creates a new PerformanceResourceTiming entry in the Resource Timeline. A
PerformanceResourceTiming is a subclass of PerformanceEntry whose
performanceEntry.entryType is always 'resource'. Performance resources
are used to mark moments in the Resource Timeline.
The created PerformanceMark entry is put in the global Resource Timeline
and can be queried with performance.getEntries,
performance.getEntriesByName, and performance.getEntriesByType. When the
observation is performed, the entries should be cleared from the global
Performance Timeline manually with performance.clearResourceTimings.
performance.measure(name, startMarkOrOptions?, endMark?): void
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, an error is thrown.
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.
The created PerformanceMeasure entry is put in the global Performance Timeline
and can be queried with performance.getEntries,
performance.getEntriesByName, and performance.getEntriesByType. When the
observation is performed, the entries should be cleared from the global
Performance Timeline manually with performance.clearMeasures.
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
History
performance object as the receiver.performance.now(): number
numberReturns the current high resolution millisecond timestamp, where 0 represents
the start of the current node process.
performance.setResourceTimingBufferSize
History
performance object as the receiver.performance.setResourceTimingBufferSize(maxSize): void
Sets the global performance resource timing buffer size to the specified number of "resource" type performance entry objects.
By default the max buffer size is set to 250.
numberThe timeOrigin specifies the high resolution millisecond timestamp at
which the current node process began, measured in Unix time.
performance.timerify(fn, options?): void
FunctionObjectRecordableHistogramperf_hooks.createHistogram() that will record runtime durations in
nanoseconds.This is an alias of perf_hooks.timerify().
This property is an extension by Node.js. It is not available in Web browsers.
performance.toJSON
History
performance object as the receiver.performance.toJSON(): void
An object which is JSON representation of the performance object. It
is similar to window.performance.toJSON in browsers.
The 'resourcetimingbufferfull' event is fired when the global performance
resource timing buffer is full. Adjust resource timing buffer size with
performance.setResourceTimingBufferSize() or clear the buffer with
performance.clearResourceTimings() in the event listener to allow
more entries to be added to the performance timeline buffer.
The constructor of this class is not exposed to users directly.
numberThe total number of milliseconds elapsed for this entry. This value will not be meaningful for all Performance Entry types.
stringThe type of the performance entry. It may be one of:
'dns'(Node.js only)'function'(Node.js only)'gc'(Node.js only)'http2'(Node.js only)'http'(Node.js only)'mark'(available on the Web)'measure'(available on the Web)'net'(Node.js only)'node'(Node.js only)'resource'(available on the Web)
stringThe name of the performance entry.
numberThe high resolution millisecond timestamp marking the starting time of the Performance Entry.
class PerformanceMark extends PerformanceEntry
Exposes marks created via the Performance.mark() method.
performanceMark.detail
History
PerformanceMark object as the receiver.anyAdditional detail specified when creating with Performance.mark() method.
class PerformanceMeasure extends PerformanceEntry
Exposes measures created via the Performance.measure() method.
The constructor of this class is not exposed to users directly.
performanceMeasure.detail
History
PerformanceMeasure object as the receiver.anyAdditional detail specified when creating with Performance.measure() method.
class PerformanceNodeEntry extends PerformanceEntry
This class is an extension by Node.js. It is not available in Web browsers.
Provides detailed Node.js timing data.
The constructor of this class is not exposed to users directly.
performanceNodeEntry.detail
History
PerformanceNodeEntry object as the receiver.anyAdditional detail specific to the entryType.
performanceNodeEntry.flags
History
performanceNodeEntry.detail instead.numberWhen performanceEntry.entryType is equal to 'gc', the performance.flags
property contains additional information about garbage collection operation.
The value may be one of:
perf_hooks.constants.NODE_PERFORMANCE_GC_FLAGS_NOperf_hooks.constants.NODE_PERFORMANCE_GC_FLAGS_CONSTRUCT_RETAINEDperf_hooks.constants.NODE_PERFORMANCE_GC_FLAGS_FORCEDperf_hooks.constants.NODE_PERFORMANCE_GC_FLAGS_SYNCHRONOUS_PHANTOM_PROCESSINGperf_hooks.constants.NODE_PERFORMANCE_GC_FLAGS_ALL_AVAILABLE_GARBAGEperf_hooks.constants.NODE_PERFORMANCE_GC_FLAGS_ALL_EXTERNAL_MEMORYperf_hooks.constants.NODE_PERFORMANCE_GC_FLAGS_SCHEDULE_IDLE
performanceNodeEntry.kind
History
performanceNodeEntry.detail instead.numberWhen 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:
perf_hooks.constants.NODE_PERFORMANCE_GC_MAJORperf_hooks.constants.NODE_PERFORMANCE_GC_MINORperf_hooks.constants.NODE_PERFORMANCE_GC_MINOR_MARK_SWEEPperf_hooks.constants.NODE_PERFORMANCE_GC_INCREMENTALperf_hooks.constants.NODE_PERFORMANCE_GC_WEAKCB
When performanceEntry.type is equal to 'gc', the
performanceNodeEntry.detail property will be an Object with two properties:
numbernumberWhen performanceEntry.type is equal to 'http', the
performanceNodeEntry.detail property will be an Object containing
additional information.
If performanceEntry.name is equal to HttpClient, the detail
will contain the following properties: req, res. And the req property
will be an Object containing method, url, headers, the res property
will be an Object containing statusCode, statusMessage, headers.
If performanceEntry.name is equal to HttpRequest, the detail
will contain the following properties: req, res. And the req property
will be an Object containing method, url, headers, the res property
will be an Object containing statusCode, statusMessage, headers.
This could add additional memory overhead and should only be used for diagnostic purposes, not left turned on in production by default.
When performanceEntry.type is equal to 'http2', the
performanceNodeEntry.detail property will be an Object containing
additional performance information.
If performanceEntry.name is equal to Http2Stream, the detail
will contain the following properties:
numberDATA frame bytes received for this
Http2Stream.numberDATA frame bytes sent for this
Http2Stream.numberHttp2StreamnumberPerformanceEntry startTime and the reception of the first DATA frame.numberPerformanceEntry startTime and sending of the first DATA frame.numberPerformanceEntry startTime and the reception of the first header.If performanceEntry.name is equal to Http2Session, the detail will
contain the following properties:
numberHttp2Session.numberHttp2Session.numberHttp2Session.numberHttp2Session.numberHttp2Session.numberPING frame and the reception of its acknowledgment. Only present if
a PING frame has been sent on the Http2Session.numberHttp2Stream instances.numberHttp2Stream instances processed by
the Http2Session.string'server' or 'client' to identify the type of
Http2Session.When performanceEntry.type is equal to 'function', the
performanceNodeEntry.detail property will be an Array listing
the input arguments to the timed function.
When performanceEntry.type is equal to 'net', the
performanceNodeEntry.detail property will be an Object containing
additional information.
If performanceEntry.name is equal to connect, the detail
will contain the following properties: host, port.
When performanceEntry.type is equal to 'dns', the
performanceNodeEntry.detail property will be an Object containing
additional information.
If performanceEntry.name is equal to lookup, the detail
will contain the following properties: hostname, family, hints, verbatim,
addresses.
If performanceEntry.name is equal to lookupService, the detail will
contain the following properties: host, port, hostname, service.
If performanceEntry.name is equal to queryxxx or getHostByAddr, the detail will
contain the following properties: host, ttl, result. The value of result is
same as the result of queryxxx or getHostByAddr.
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.
numberThe 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.
numberThe high resolution millisecond timestamp at which the Node.js environment was initialized.
numberThe 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.
numberThe 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.
numberThe 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.
numberThe high resolution millisecond timestamp at which the Node.js process was initialized.
This is a wrapper to the uv_metrics_info function.
It returns the current set of event loop metrics.
It is recommended to use this property inside a function whose execution was
scheduled using setImmediate to avoid collecting metrics before finishing all
operations scheduled during the current loop iteration.
const { performance } = require('node:perf_hooks'); setImmediate(() => { console.log(performance.nodeTiming.uvMetricsInfo); });
import { performance } from 'node:perf_hooks'; setImmediate(() => { console.log(performance.nodeTiming.uvMetricsInfo); });
numberThe high resolution millisecond timestamp at which the V8 platform was initialized.
class PerformanceResourceTiming extends PerformanceEntry
Provides detailed network timing data regarding the loading of an application's resources.
The constructor of this class is not exposed to users directly.
performanceResourceTiming.workerStart
History
PerformanceResourceTiming object as the receiver.numberThe high resolution millisecond timestamp at immediately before dispatching
the fetch request. If the resource is not intercepted by a worker the property
will always return 0.
performanceResourceTiming.redirectStart
History
PerformanceResourceTiming object as the receiver.numberThe high resolution millisecond timestamp that represents the start time of the fetch which initiates the redirect.
performanceResourceTiming.redirectEnd
History
PerformanceResourceTiming object as the receiver.numberThe high resolution millisecond timestamp that will be created immediately after receiving the last byte of the response of the last redirect.
performanceResourceTiming.fetchStart
History
PerformanceResourceTiming object as the receiver.numberThe high resolution millisecond timestamp immediately before the Node.js starts to fetch the resource.
performanceResourceTiming.domainLookupStart
History
PerformanceResourceTiming object as the receiver.numberThe high resolution millisecond timestamp immediately before the Node.js starts the domain name lookup for the resource.
performanceResourceTiming.domainLookupEnd
History
PerformanceResourceTiming object as the receiver.numberThe high resolution millisecond timestamp representing the time immediately after the Node.js finished the domain name lookup for the resource.
performanceResourceTiming.connectStart
History
PerformanceResourceTiming object as the receiver.numberThe high resolution millisecond timestamp representing the time immediately before Node.js starts to establish the connection to the server to retrieve the resource.
performanceResourceTiming.connectEnd
History
PerformanceResourceTiming object as the receiver.numberThe high resolution millisecond timestamp representing the time immediately after Node.js finishes establishing the connection to the server to retrieve the resource.
performanceResourceTiming.secureConnectionStart
History
PerformanceResourceTiming object as the receiver.numberThe high resolution millisecond timestamp representing the time immediately before Node.js starts the handshake process to secure the current connection.
performanceResourceTiming.requestStart
History
PerformanceResourceTiming object as the receiver.numberThe high resolution millisecond timestamp representing the time immediately before Node.js receives the first byte of the response from the server.
performanceResourceTiming.responseEnd
History
PerformanceResourceTiming object as the receiver.numberThe high resolution millisecond timestamp representing the time immediately after Node.js receives the last byte of the resource or immediately before the transport connection is closed, whichever comes first.
performanceResourceTiming.transferSize
History
PerformanceResourceTiming object as the receiver.numberA number representing the size (in octets) of the fetched resource. The size includes the response header fields plus the response payload body.
performanceResourceTiming.encodedBodySize
History
PerformanceResourceTiming object as the receiver.numberA number representing the size (in octets) received from the fetch (HTTP or cache), of the payload body, before removing any applied content-codings.
performanceResourceTiming.decodedBodySize
History
PerformanceResourceTiming object as the receiver.numberA number representing the size (in octets) received from the fetch (HTTP or cache), of the message body, after removing any applied content-codings.
performanceResourceTiming.toJSON
History
PerformanceResourceTiming object as the receiver.performanceResourceTiming.toJSON(): void
Returns a object that is the JSON representation of the
PerformanceResourceTiming object
string[]Get supported types.
new PerformanceObserver(callback): PerformanceObserver
FunctionPerformanceObserverPerformanceObserver objects provide notifications when new
PerformanceEntry instances have been added to the Performance Timeline.
import { performance, PerformanceObserver } from 'node:perf_hooks'; const obs = new PerformanceObserver((list, observer) => { console.log(list.getEntries()); performance.clearMarks(); performance.clearMeasures(); observer.disconnect(); }); obs.observe({ entryTypes: ['mark'], buffered: true }); performance.mark('test');
const { performance, PerformanceObserver, } = require('node:perf_hooks'); const obs = new PerformanceObserver((list, observer) => { console.log(list.getEntries()); performance.clearMarks(); performance.clearMeasures(); 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(): void
Disconnects the PerformanceObserver instance from all notifications.
performanceObserver.observe(options): void
ObjectstringPerformanceEntry type. Must not be given
if entryTypes is already specified.string[]PerformanceEntry instances the observer is interested in. If not
provided an error will be thrown.booleanPerformanceEntry buffered entries. If false, only
PerformanceEntrys created after the time point are sent to the
observer callback. Default: false.Subscribes the PerformanceObserver instance to notifications of new
PerformanceEntry instances identified either by options.entryTypes
or options.type:
import { performance, PerformanceObserver } from 'node:perf_hooks'; const obs = new PerformanceObserver((list, observer) => { // Called once asynchronously. `list` contains three items. }); obs.observe({ type: 'mark' }); for (let n = 0; n < 3; n++) performance.mark(`test${n}`);
const { performance, PerformanceObserver, } = require('node:perf_hooks'); const obs = new PerformanceObserver((list, observer) => { // Called once asynchronously. `list` contains three items. }); obs.observe({ type: 'mark' }); for (let n = 0; n < 3; n++) performance.mark(`test${n}`);
performanceObserver.takeRecords(): PerformanceEntry[]
PerformanceEntry[]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(): PerformanceEntry[]
PerformanceEntry[]Returns a list of PerformanceEntry objects in chronological order
with respect to performanceEntry.startTime.
import { performance, PerformanceObserver } from 'node:perf_hooks'; const obs = new PerformanceObserver((perfObserverList, observer) => { console.log(perfObserverList.getEntries()); /** * [ * PerformanceEntry { * name: 'test', * entryType: 'mark', * startTime: 81.465639, * duration: 0, * detail: null * }, * PerformanceEntry { * name: 'meow', * entryType: 'mark', * startTime: 81.860064, * duration: 0, * detail: null * } * ] */ performance.clearMarks(); performance.clearMeasures(); observer.disconnect(); }); obs.observe({ type: 'mark' }); performance.mark('test'); performance.mark('meow');
const { performance, PerformanceObserver, } = require('node:perf_hooks'); const obs = new PerformanceObserver((perfObserverList, observer) => { console.log(perfObserverList.getEntries()); /** * [ * PerformanceEntry { * name: 'test', * entryType: 'mark', * startTime: 81.465639, * duration: 0, * detail: null * }, * PerformanceEntry { * name: 'meow', * entryType: 'mark', * startTime: 81.860064, * duration: 0, * detail: null * } * ] */ performance.clearMarks(); performance.clearMeasures(); observer.disconnect(); }); obs.observe({ type: 'mark' }); performance.mark('test'); performance.mark('meow');
performanceObserverEntryList.getEntriesByName(name, type?): PerformanceEntry[]
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.
import { performance, PerformanceObserver } from 'node:perf_hooks'; const obs = new PerformanceObserver((perfObserverList, observer) => { console.log(perfObserverList.getEntriesByName('meow')); /** * [ * PerformanceEntry { * name: 'meow', * entryType: 'mark', * startTime: 98.545991, * duration: 0, * detail: null * } * ] */ console.log(perfObserverList.getEntriesByName('nope')); // [] console.log(perfObserverList.getEntriesByName('test', 'mark')); /** * [ * PerformanceEntry { * name: 'test', * entryType: 'mark', * startTime: 63.518931, * duration: 0, * detail: null * } * ] */ console.log(perfObserverList.getEntriesByName('test', 'measure')); // [] performance.clearMarks(); performance.clearMeasures(); observer.disconnect(); }); obs.observe({ entryTypes: ['mark', 'measure'] }); performance.mark('test'); performance.mark('meow');
const { performance, PerformanceObserver, } = require('node:perf_hooks'); const obs = new PerformanceObserver((perfObserverList, observer) => { console.log(perfObserverList.getEntriesByName('meow')); /** * [ * PerformanceEntry { * name: 'meow', * entryType: 'mark', * startTime: 98.545991, * duration: 0, * detail: null * } * ] */ console.log(perfObserverList.getEntriesByName('nope')); // [] console.log(perfObserverList.getEntriesByName('test', 'mark')); /** * [ * PerformanceEntry { * name: 'test', * entryType: 'mark', * startTime: 63.518931, * duration: 0, * detail: null * } * ] */ console.log(perfObserverList.getEntriesByName('test', 'measure')); // [] performance.clearMarks(); performance.clearMeasures(); observer.disconnect(); }); obs.observe({ entryTypes: ['mark', 'measure'] }); performance.mark('test'); performance.mark('meow');
performanceObserverEntryList.getEntriesByType(type): PerformanceEntry[]
stringPerformanceEntry[]Returns a list of PerformanceEntry objects in chronological order
with respect to performanceEntry.startTime whose performanceEntry.entryType
is equal to type.
import { performance, PerformanceObserver } from 'node:perf_hooks'; const obs = new PerformanceObserver((perfObserverList, observer) => { console.log(perfObserverList.getEntriesByType('mark')); /** * [ * PerformanceEntry { * name: 'test', * entryType: 'mark', * startTime: 55.897834, * duration: 0, * detail: null * }, * PerformanceEntry { * name: 'meow', * entryType: 'mark', * startTime: 56.350146, * duration: 0, * detail: null * } * ] */ performance.clearMarks(); performance.clearMeasures(); observer.disconnect(); }); obs.observe({ type: 'mark' }); performance.mark('test'); performance.mark('meow');
const { performance, PerformanceObserver, } = require('node:perf_hooks'); const obs = new PerformanceObserver((perfObserverList, observer) => { console.log(perfObserverList.getEntriesByType('mark')); /** * [ * PerformanceEntry { * name: 'test', * entryType: 'mark', * startTime: 55.897834, * duration: 0, * detail: null * }, * PerformanceEntry { * name: 'meow', * entryType: 'mark', * startTime: 56.350146, * duration: 0, * detail: null * } * ] */ performance.clearMarks(); performance.clearMeasures(); observer.disconnect(); }); obs.observe({ type: 'mark' }); performance.mark('test'); performance.mark('meow');
perf_hooks.createHistogram(options?): RecordableHistogram
Object1.lowest.
Default: Number.MAX_SAFE_INTEGER.number1 and 5. Default: 3.numberhistogram.ewmaMean and histogram.ewmaStddev. After halfLife
recordings, a value's influence has decayed to 50%. Default: 0
(disabled).numberhalfLife, the histogram tracks a smoothed error rate for values
exceeding this threshold, accessible via histogram.ewmaErrorRate and
histogram.burnRate(). Default: 0 (disabled).RecordableHistogramReturns a RecordableHistogram.
perf_hooks.eventLoopUtilization(utilization1?, utilization2?): Object
The eventLoopUtilization() function 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.
import { eventLoopUtilization } from 'node:perf_hooks'; import { spawnSync } from 'node:child_process'; setImmediate(() => { const elu = eventLoopUtilization(); spawnSync('sleep', ['5']); console.log(eventLoopUtilization(elu).utilization); });
const { eventLoopUtilization } = require('node:perf_hooks'); const { spawnSync } = require('node: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.
perf_hooks.monitorEventLoopDelay
History
samplePerIteration option.perf_hooks.monitorEventLoopDelay(options?): ELDHistogram
ObjectELDHistogramThis 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.
By default, the histogram is updated by a timer using the configured
resolution. When samplePerIteration is true, samples are taken once per
event loop iteration using uv_prepare_t and uv_check_t hooks. In that mode,
the histogram does not keep the loop alive or force additional iterations when
the application is idle.
The two sampling modes produce significantly different results and should not
be compared directly.
import { monitorEventLoopDelay } from 'node: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));
const { monitorEventLoopDelay } = require('node: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));
perf_hooks.timerify(fn, options?): void
FunctionObjectRecordableHistogramperf_hooks.createHistogram() that will record runtime durations in
nanoseconds.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.
import { timerify, performance, PerformanceObserver } from 'node:perf_hooks'; function someFunction() { console.log('hello world'); } const wrapped = timerify(someFunction); const obs = new PerformanceObserver((list) => { console.log(list.getEntries()[0].duration); performance.clearMarks(); performance.clearMeasures(); obs.disconnect(); }); obs.observe({ entryTypes: ['function'] }); // A performance timeline entry will be created wrapped();
const { timerify, performance, PerformanceObserver, } = require('node:perf_hooks'); function someFunction() { console.log('hello world'); } const wrapped = timerify(someFunction); const obs = new PerformanceObserver((list) => { console.log(list.getEntries()[0].duration); performance.clearMarks(); performance.clearMeasures(); obs.disconnect(); }); obs.observe({ entryTypes: ['function'] }); // A performance timeline entry will be created wrapped();
If the wrapped function returns a promise, a finally handler will be attached to the promise and the duration will be reported once the finally handler is invoked.
numberThe number of samples recorded by the histogram.
bigintThe number of samples recorded by the histogram.
histogram.ccdf(value): number
Returns the complementary cumulative distribution function (CCDF) value
for the given value, representing the probability that a recorded value
will exceed value. Equivalent to 1 - histogram.cdf(value).
histogram.cdf(value): number
Returns the cumulative distribution function (CDF) value for the given
value, representing the probability that a recorded value will be less
than or equal to value. This is the inverse operation of
histogram.percentile().
histogram.cliffsD(other): number
Computes Cliff's delta, a non-parametric effect size measure. Returns
the probability that a random value from this histogram exceeds a random
value from other, minus the reverse probability. A value of 1 means every
value in this histogram exceeds every value in other; -1 means the
opposite; 0 means no tendency in either direction.
histogram.cohensD(other): number
Computes Cohen's d effect size, the standardized difference between the
means of this histogram and other, using the pooled standard deviation.
Positive values indicate this histogram has a higher mean. By convention,
|d| < 0.2 is a small effect, 0.5 is medium, and 0.8 or greater is large.
Both histograms must have at least 2 recorded values; otherwise returns 0.
histogram.countAt(value): number
Returns the number of recorded values that fall within the equivalent value range of the given value.
numberThe number of times the event loop delay exceeded the maximum 1 hour event loop delay threshold.
bigintThe number of times the event loop delay exceeded the maximum 1 hour event loop delay threshold.
numberThe exponentially weighted moving average of recorded values. Only active
when the histogram was created with a halfLife option greater than 0.
Returns 0 when EWMA is disabled or no values have been recorded.
numberThe exponentially weighted moving standard deviation. Only active when the
histogram was created with a halfLife option greater than 0. Returns 0
when EWMA is disabled or no values have been recorded.
numberThe EWMA-smoothed probability of a recorded value exceeding the configured
threshold. Only active when the histogram was created with both halfLife
and threshold options. Returns 0 when not enabled or no values have been
recorded.
histogram.burnRate(sloTarget): number
Returns the SLO burn rate: ewmaErrorRate / (1 - sloTarget). A burn rate
of 1 means the error budget will be exactly exhausted over the SLO window.
A burn rate greater than 1 means it is being consumed faster than allowed.
Requires the histogram to have been created with both halfLife and
threshold options.
const { createHistogram } = require('node:perf_hooks'); // Track latency with a 200ms SLO threshold, half-life of 100 samples const h = createHistogram({ halfLife: 100, threshold: 200_000_000 }); // ... record latency values ... // Check burn rate against a 99.9% SLO const rate = h.burnRate(0.999); if (rate > 1) { console.log(`SLO burn rate: ${rate.toFixed(2)}x — error budget depleting`); }
histogram.ksTest(other): number
Computes the Kolmogorov-Smirnov test statistic comparing this histogram's
distribution to other. A value of 0 indicates identical distributions;
values close to 1 indicate completely disjoint distributions. Useful for
detecting performance regressions by comparing before/after histograms.
numberThe excess kurtosis of the recorded values. Measures the heaviness of the distribution's tails relative to a normal distribution. Positive values indicate heavier tails (more extreme outliers); negative values indicate lighter tails.
histogram.linearBuckets(stepSize): Map
Returns the histogram data rebucketed into linearly-spaced intervals
of stepSize. Useful for visualization and export.
histogram.logBuckets(firstBucket, base): Map
Returns the histogram data rebucketed into logarithmically-spaced
intervals, where each bucket's width is multiplied by base.
Useful for visualization and export.
histogram.mannWhitneyTest(other): Object
Performs a Mann-Whitney U test comparing whether this histogram tends to
produce larger or smaller values than other. Unlike welchTest(), this is a
non-parametric test that makes no assumptions about the shape of the
distributions. Uses the normal approximation with tie correction for the
p-value.
numberThe maximum recorded event loop delay.
bigintThe maximum recorded event loop delay.
numberThe mean of the recorded event loop delays.
numberThe minimum recorded event loop delay.
bigintThe minimum recorded event loop delay.
histogram.percentile(percentile): number
Returns the value at the given percentile.
histogram.percentileBigInt(percentile): bigint
Returns the value at the given percentile.
histogram.percentileCI(percentile, options?): Object
Returns a confidence interval for the given percentile using the exact
binomial method. With fewer samples, the interval will be wider, reflecting
the greater uncertainty in the percentile estimate. Requires at least 2
recorded values; with fewer than 2, lower and upper will equal value.
const { createHistogram } = require('node:perf_hooks'); const h = createHistogram(); for (let i = 0; i < 1000; i++) { h.record(Math.floor(Math.random() * 100)); } const ci = h.percentileCI(99); console.log(ci.value); // The p99 point estimate console.log(ci.lower); // The lower bound (95% confidence) console.log(ci.upper); // The upper bound (95% confidence)
MapReturns a Map object detailing the accumulated percentile distribution.
MapReturns a Map object detailing the accumulated percentile distribution.
histogram.percentilesAt(percentiles): Map
Returns the values at the specified percentiles, computed in a single
efficient pass over the histogram data. More efficient than calling
histogram.percentile() multiple times.
histogram.reset(): void
Resets the collected histogram data.
numberThe skewness of the recorded values. Measures the asymmetry of the distribution. A positive value indicates a right-skewed distribution (longer right tail, common for latency data); a negative value indicates a left-skewed distribution.
numberThe standard deviation of the recorded event loop delays.
histogram.welchTest(other, options?): Object
HistogramPerforms Welch's t-test comparing the means of this histogram and other.
The p-value indicates the probability of observing a difference at least this
extreme under the null hypothesis that the two distributions have the same
mean. Both histograms must have at least 2 recorded values; otherwise the
result has pValue 1 and tStatistic 0.
A Histogram that records event loop delay, returned by
perf_hooks.monitorEventLoopDelay().
histogram.disable(): boolean
booleanDisables event loop delay sampling. Returns true if sampling was
stopped, false if it was already stopped.
histogram.enable(): boolean
booleanEnables event loop delay sampling. Returns true if sampling was
started, false if it was already started.
histogram[Symbol.dispose](): void
Disables event loop delay sampling when the histogram is disposed.
const { monitorEventLoopDelay } = require('node:perf_hooks'); { using hist = monitorEventLoopDelay({ resolution: 20 }); hist.enable(); // The histogram will be disabled when the block is exited. }
ELDHistogram instances can be cloned via MessagePort. On the receiving end,
the histogram is cloned as a plain Histogram object that does not implement
the enable() and disable() methods.
histogram.add(other): void
RecordableHistogramAdds the values from other to this histogram.
histogram.record(val): void
histogram.recordDelta(): void
Calculates the amount of time (in nanoseconds) that has passed since the
previous call to recordDelta() and records that amount in the histogram.
histogram.recordCorrected(val, expectedInterval): void
Records a value with coordinated omission correction. When a system stall
prevents timely recording, this method backfills intermediate values at
expectedInterval steps between the previously recorded value and val.
This compensates for measurement gaps that would otherwise underrepresent
latency.
histogram.subtract(other): void
RecordableHistogramSubtracts the values of other from this histogram. Both histograms should
have compatible configurations. Bucket counts that would become negative
are clamped to zero.
The Histogram class provides statistical analysis methods useful for
performance monitoring, SLO enforcement, and regression detection.
const { createHistogram } = require('node:perf_hooks'); const h = createHistogram(); // Simulate a right-skewed latency distribution for (let i = 0; i < 1000; i++) { h.record(Math.ceil(Math.random() * 100)); } // Add some outliers for (let i = 0; i < 10; i++) { h.record(500 + Math.ceil(Math.random() * 500)); } console.log('Skewness:', h.skewness.toFixed(4)); // Positive = right-skewed console.log('Kurtosis:', h.kurtosis.toFixed(4)); // Positive = heavy tails
const { createHistogram } = require('node:perf_hooks'); const latency = createHistogram(); // Record request latencies (in nanoseconds)... // "What fraction of requests complete within 100ms?" const withinSLO = latency.cdf(100_000_000); console.log(`${(withinSLO * 100).toFixed(1)}% of requests within SLO`); // "What fraction of requests exceed 500ms?" const violating = latency.ccdf(500_000_000); console.log(`${(violating * 100).toFixed(1)}% of requests violating SLO`);
const { createHistogram } = require('node:perf_hooks'); // Track latency with EWMA (half-life 100 samples) and a 200ms SLO threshold const latency = createHistogram({ halfLife: 100, threshold: 200_000_000, // 200ms in nanoseconds }); // Record request latencies... // Smoothed error rate: probability of exceeding the threshold console.log(`Error rate: ${(latency.ewmaErrorRate * 100).toFixed(2)}%`); // Burn rate against a 99.9% SLO // >1 means the error budget is depleting faster than allowed const rate = latency.burnRate(0.999); console.log(`Burn rate: ${rate.toFixed(2)}x`); // EWMA mean and stddev track the smoothed latency console.log(`EWMA latency: ${latency.ewmaMean.toFixed(0)}ns`); console.log(`EWMA stddev: ${latency.ewmaStddev.toFixed(0)}ns`);
const { createHistogram } = require('node:perf_hooks'); const baseline = createHistogram(); const current = createHistogram(); // Record baseline and current latencies... // D-statistic: 0 = identical, 1 = completely different const d = baseline.ksTest(current); if (d > 0.1) { console.log(`Possible regression detected (D=${d.toFixed(4)})`); }
const { createHistogram } = require('node:perf_hooks'); const h = createHistogram(); // Record values... // Efficiently query common monitoring percentiles in one pass const p = h.percentilesAt([50, 75, 90, 95, 99, 99.9]); console.log('p50:', p.get(50)); console.log('p99:', p.get(99));
const { createHistogram } = require('node:perf_hooks'); const total = createHistogram(); const snapshot = createHistogram(); // Record values into total... // Periodically snapshot for "last interval" analysis: snapshot.add(total); // Later, take a new snapshot and diff: const newSnapshot = createHistogram(); newSnapshot.add(total); newSnapshot.subtract(snapshot); // newSnapshot now contains only the values recorded since the last snapshot console.log('Recent p99:', newSnapshot.percentile(99));
const { createHistogram } = require('node:perf_hooks'); const baseline = createHistogram(); const candidate = createHistogram(); // Record operation rates from the old and new builds... const result = baseline.welchTest(candidate); const improvement = ((candidate.mean - baseline.mean) / baseline.mean * 100); console.log(`Improvement: ${improvement.toFixed(2)}%`); console.log(`p-value: ${result.pValue.toFixed(6)}`); console.log(`95% CI: [${result.confidenceInterval.lower.toFixed(2)}, ` + `${result.confidenceInterval.upper.toFixed(2)}]`); if (result.pValue < 0.05) { const d = baseline.cohensD(candidate); console.log(`Statistically significant (Cohen's d = ${d.toFixed(4)})`); }
const { createHistogram } = require('node:perf_hooks'); const before = createHistogram(); const after = createHistogram(); // Record latencies before and after a change... const delta = before.cliffsD(after); // A delta > 0: before tends to produce larger values (improvement) // A delta < 0: after tends to produce larger values (regression) console.log(`Cliff's delta: ${delta.toFixed(4)}`);
The following example uses the Async Hooks and Performance APIs to measure the actual duration of a Timeout operation (including the amount of time it took to execute the callback).
import { createHook } from 'node:async_hooks'; import { performance, PerformanceObserver } from 'node:perf_hooks'; const set = new Set(); const hook = createHook({ init(id, type) { if (type === 'Timeout') { performance.mark(`Timeout-${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(); performance.clearMeasures(); observer.disconnect(); }); obs.observe({ entryTypes: ['measure'], buffered: true }); setTimeout(() => {}, 1000);
const async_hooks = require('node:async_hooks'); const { performance, PerformanceObserver, } = require('node: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-${id}-Destroy`); } }, }); hook.enable(); const obs = new PerformanceObserver((list, observer) => { console.log(list.getEntries()[0]); performance.clearMarks(); performance.clearMeasures(); observer.disconnect(); }); obs.observe({ entryTypes: ['measure'] }); setTimeout(() => {}, 1000);
The following example measures the duration of require() operations to load
dependencies:
import { performance, PerformanceObserver } from 'node:perf_hooks'; // Activate the observer const obs = new PerformanceObserver((list) => { const entries = list.getEntries(); entries.forEach((entry) => { console.log(`import('${entry[0]}')`, entry.duration); }); performance.clearMarks(); performance.clearMeasures(); obs.disconnect(); }); obs.observe({ entryTypes: ['function'], buffered: true }); const timedImport = performance.timerify(async (module) => { return await import(module); }); await timedImport('some-module');
const { performance, PerformanceObserver, } = require('node:perf_hooks'); const mod = require('node: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('${entry[0]}')`, entry.duration); }); performance.clearMarks(); performance.clearMeasures(); obs.disconnect(); }); obs.observe({ entryTypes: ['function'] }); require('some-module');
The following example is used to trace the time spent by HTTP client
(OutgoingMessage) and HTTP request (IncomingMessage). For HTTP client,
it means the time interval between starting the request and receiving the
response, and for HTTP request, it means the time interval between receiving
the request and sending the response:
import { PerformanceObserver } from 'node:perf_hooks'; import { createServer, get } from 'node:http'; const obs = new PerformanceObserver((items) => { items.getEntries().forEach((item) => { console.log(item); }); }); obs.observe({ entryTypes: ['http'] }); const PORT = 8080; createServer((req, res) => { res.end('ok'); }).listen(PORT, () => { get(`http://127.0.0.1:${PORT}`); });
const { PerformanceObserver } = require('node:perf_hooks'); const http = require('node:http'); const obs = new PerformanceObserver((items) => { items.getEntries().forEach((item) => { console.log(item); }); }); obs.observe({ entryTypes: ['http'] }); const PORT = 8080; http.createServer((req, res) => { res.end('ok'); }).listen(PORT, () => { http.get(`http://127.0.0.1:${PORT}`); });
import { PerformanceObserver } from 'node:perf_hooks'; import { connect, createServer } from 'node:net'; const obs = new PerformanceObserver((items) => { items.getEntries().forEach((item) => { console.log(item); }); }); obs.observe({ entryTypes: ['net'] }); const PORT = 8080; createServer((socket) => { socket.destroy(); }).listen(PORT, () => { connect(PORT); });
const { PerformanceObserver } = require('node:perf_hooks'); const net = require('node:net'); const obs = new PerformanceObserver((items) => { items.getEntries().forEach((item) => { console.log(item); }); }); obs.observe({ entryTypes: ['net'] }); const PORT = 8080; net.createServer((socket) => { socket.destroy(); }).listen(PORT, () => { net.connect(PORT); });
import { PerformanceObserver } from 'node:perf_hooks'; import { lookup, promises } from 'node:dns'; const obs = new PerformanceObserver((items) => { items.getEntries().forEach((item) => { console.log(item); }); }); obs.observe({ entryTypes: ['dns'] }); lookup('localhost', () => {}); promises.resolve('localhost');
const { PerformanceObserver } = require('node:perf_hooks'); const dns = require('node:dns'); const obs = new PerformanceObserver((items) => { items.getEntries().forEach((item) => { console.log(item); }); }); obs.observe({ entryTypes: ['dns'] }); dns.lookup('localhost', () => {}); dns.promises.resolve('localhost');