{ "type": "module", "source": "doc/api/async_hooks.md", "modules": [ { "textRaw": "Async Hooks", "name": "async_hooks", "introduced_in": "v8.1.0", "stability": 1, "stabilityText": "Experimental", "desc": "
The async_hooks
module provides an API to register callbacks tracking the\nlifetime of asynchronous resources created inside a Node.js application.\nIt can be accessed using:
const async_hooks = require('async_hooks');\n
",
"modules": [
{
"textRaw": "Terminology",
"name": "terminology",
"desc": "An asynchronous resource represents an object with an associated callback.\nThis callback may be called multiple times, for example, the 'connection'
\nevent in net.createServer()
, or just a single time like in fs.open()
.\nA resource can also be closed before the callback is called. AsyncHook
does\nnot explicitly distinguish between these different cases but will represent them\nas the abstract concept that is a resource.
If Worker
s are used, each thread has an independent async_hooks
\ninterface, and each thread will use a new set of async IDs.
Following is a simple overview of the public API.
\nconst async_hooks = require('async_hooks');\n\n// Return the ID of the current execution context.\nconst eid = async_hooks.executionAsyncId();\n\n// Return the ID of the handle responsible for triggering the callback of the\n// current execution scope to call.\nconst tid = async_hooks.triggerAsyncId();\n\n// Create a new AsyncHook instance. All of these callbacks are optional.\nconst asyncHook =\n async_hooks.createHook({ init, before, after, destroy, promiseResolve });\n\n// Allow callbacks of this AsyncHook instance to call. This is not an implicit\n// action after running the constructor, and must be explicitly run to begin\n// executing callbacks.\nasyncHook.enable();\n\n// Disable listening for new asynchronous events.\nasyncHook.disable();\n\n//\n// The following are the callbacks that can be passed to createHook().\n//\n\n// init is called during object construction. The resource may not have\n// completed construction when this callback runs, therefore all fields of the\n// resource referenced by \"asyncId\" may not have been populated.\nfunction init(asyncId, type, triggerAsyncId, resource) { }\n\n// Before is called just before the resource's callback is called. It can be\n// called 0-N times for handles (e.g. TCPWrap), and will be called exactly 1\n// time for requests (e.g. FSReqCallback).\nfunction before(asyncId) { }\n\n// After is called just after the resource's callback has finished.\nfunction after(asyncId) { }\n\n// Destroy is called when an AsyncWrap instance is destroyed.\nfunction destroy(asyncId) { }\n\n// promiseResolve is called only for promise resources, when the\n// `resolve` function passed to the `Promise` constructor is invoked\n// (either directly or through other means of resolving a promise).\nfunction promiseResolve(asyncId) { }\n
",
"modules": [
{
"textRaw": "`async_hooks.createHook(callbacks)`",
"name": "`async_hooks.createhook(callbacks)`",
"meta": {
"added": [
"v8.1.0"
],
"changes": []
},
"desc": "callbacks
<Object> The Hook Callbacks to register
init
<Function> The init
callback.before
<Function> The before
callback.after
<Function> The after
callback.destroy
<Function> The destroy
callback.promiseResolve
<Function> The promiseResolve
callback.Registers functions to be called for different lifetime events of each async\noperation.
\nThe callbacks init()
/before()
/after()
/destroy()
are called for the\nrespective asynchronous event during a resource's lifetime.
All callbacks are optional. For example, if only resource cleanup needs to\nbe tracked, then only the destroy
callback needs to be passed. The\nspecifics of all functions that can be passed to callbacks
is in the\nHook Callbacks section.
const async_hooks = require('async_hooks');\n\nconst asyncHook = async_hooks.createHook({\n init(asyncId, type, triggerAsyncId, resource) { },\n destroy(asyncId) { }\n});\n
\nThe callbacks will be inherited via the prototype chain:
\nclass MyAsyncCallbacks {\n init(asyncId, type, triggerAsyncId, resource) { }\n destroy(asyncId) {}\n}\n\nclass MyAddedCallbacks extends MyAsyncCallbacks {\n before(asyncId) { }\n after(asyncId) { }\n}\n\nconst asyncHook = async_hooks.createHook(new MyAddedCallbacks());\n
",
"modules": [
{
"textRaw": "Error Handling",
"name": "error_handling",
"desc": "If any AsyncHook
callbacks throw, the application will print the stack trace\nand exit. The exit path does follow that of an uncaught exception, but\nall 'uncaughtException'
listeners are removed, thus forcing the process to\nexit. The 'exit'
callbacks will still be called unless the application is run\nwith --abort-on-uncaught-exception
, in which case a stack trace will be\nprinted and the application exits, leaving a core file.
The reason for this error handling behavior is that these callbacks are running\nat potentially volatile points in an object's lifetime, for example during\nclass construction and destruction. Because of this, it is deemed necessary to\nbring down the process quickly in order to prevent an unintentional abort in the\nfuture. This is subject to change in the future if a comprehensive analysis is\nperformed to ensure an exception can follow the normal control flow without\nunintentional side effects.
", "type": "module", "displayName": "Error Handling" }, { "textRaw": "Printing in AsyncHooks callbacks", "name": "printing_in_asynchooks_callbacks", "desc": "Because printing to the console is an asynchronous operation, console.log()
\nwill cause the AsyncHooks callbacks to be called. Using console.log()
or\nsimilar asynchronous operations inside an AsyncHooks callback function will thus\ncause an infinite recursion. An easy solution to this when debugging is to use a\nsynchronous logging operation such as fs.writeFileSync(file, msg, flag)
.\nThis will print to the file and will not invoke AsyncHooks recursively because\nit is synchronous.
const fs = require('fs');\nconst util = require('util');\n\nfunction debug(...args) {\n // Use a function like this one when debugging inside an AsyncHooks callback\n fs.writeFileSync('log.out', `${util.format(...args)}\\n`, { flag: 'a' });\n}\n
\nIf an asynchronous operation is needed for logging, it is possible to keep\ntrack of what caused the asynchronous operation using the information\nprovided by AsyncHooks itself. The logging should then be skipped when\nit was the logging itself that caused AsyncHooks callback to call. By\ndoing this the otherwise infinite recursion is broken.
", "type": "module", "displayName": "Printing in AsyncHooks callbacks" } ], "type": "module", "displayName": "`async_hooks.createHook(callbacks)`" }, { "textRaw": "`asyncHook.enable()`", "name": "`asynchook.enable()`", "desc": "asyncHook
.Enable the callbacks for a given AsyncHook
instance. If no callbacks are\nprovided enabling is a noop.
The AsyncHook
instance is disabled by default. If the AsyncHook
instance\nshould be enabled immediately after creation, the following pattern can be used.
const async_hooks = require('async_hooks');\n\nconst hook = async_hooks.createHook(callbacks).enable();\n
",
"type": "module",
"displayName": "`asyncHook.enable()`"
},
{
"textRaw": "`asyncHook.disable()`",
"name": "`asynchook.disable()`",
"desc": "asyncHook
.Disable the callbacks for a given AsyncHook
instance from the global pool of\nAsyncHook
callbacks to be executed. Once a hook has been disabled it will not\nbe called again until enabled.
For API consistency disable()
also returns the AsyncHook
instance.
Key events in the lifetime of asynchronous events have been categorized into\nfour areas: instantiation, before/after the callback is called, and when the\ninstance is destroyed.
", "modules": [ { "textRaw": "`init(asyncId, type, triggerAsyncId, resource)`", "name": "`init(asyncid,_type,_triggerasyncid,_resource)`", "desc": "asyncId
<number> A unique ID for the async resource.type
<string> The type of the async resource.triggerAsyncId
<number> The unique ID of the async resource in whose\nexecution context this async resource was created.resource
<Object> Reference to the resource representing the async\noperation, needs to be released during destroy.Called when a class is constructed that has the possibility to emit an\nasynchronous event. This does not mean the instance must call\nbefore
/after
before destroy
is called, only that the possibility\nexists.
This behavior can be observed by doing something like opening a resource then\nclosing it before the resource can be used. The following snippet demonstrates\nthis.
\nrequire('net').createServer().listen(function() { this.close(); });\n// OR\nclearTimeout(setTimeout(() => {}, 10));\n
\nEvery new resource is assigned an ID that is unique within the scope of the\ncurrent Node.js instance.
", "modules": [ { "textRaw": "`type`", "name": "`type`", "desc": "The type
is a string identifying the type of resource that caused\ninit
to be called. Generally, it will correspond to the name of the\nresource's constructor.
FSEVENTWRAP, FSREQCALLBACK, GETADDRINFOREQWRAP, GETNAMEINFOREQWRAP, HTTPINCOMINGMESSAGE,\nHTTPCLIENTREQUEST, JSSTREAM, PIPECONNECTWRAP, PIPEWRAP, PROCESSWRAP, QUERYWRAP,\nSHUTDOWNWRAP, SIGNALWRAP, STATWATCHER, TCPCONNECTWRAP, TCPSERVERWRAP, TCPWRAP,\nTTYWRAP, UDPSENDWRAP, UDPWRAP, WRITEWRAP, ZLIB, SSLCONNECTION, PBKDF2REQUEST,\nRANDOMBYTESREQUEST, TLSWRAP, Microtask, Timeout, Immediate, TickObject\n
\nThere is also the PROMISE
resource type, which is used to track Promise
\ninstances and asynchronous work scheduled by them.
Users are able to define their own type
when using the public embedder API.
It is possible to have type name collisions. Embedders are encouraged to use\nunique prefixes, such as the npm package name, to prevent collisions when\nlistening to the hooks.
", "type": "module", "displayName": "`type`" }, { "textRaw": "`triggerAsyncId`", "name": "`triggerasyncid`", "desc": "triggerAsyncId
is the asyncId
of the resource that caused (or \"triggered\")\nthe new resource to initialize and that caused init
to call. This is different\nfrom async_hooks.executionAsyncId()
that only shows when a resource was\ncreated, while triggerAsyncId
shows why a resource was created.
The following is a simple demonstration of triggerAsyncId
:
async_hooks.createHook({\n init(asyncId, type, triggerAsyncId) {\n const eid = async_hooks.executionAsyncId();\n fs.writeSync(\n 1, `${type}(${asyncId}): trigger: ${triggerAsyncId} execution: ${eid}\\n`);\n }\n}).enable();\n\nrequire('net').createServer((conn) => {}).listen(8080);\n
\nOutput when hitting the server with nc localhost 8080
:
TCPSERVERWRAP(5): trigger: 1 execution: 1\nTCPWRAP(7): trigger: 5 execution: 0\n
\nThe TCPSERVERWRAP
is the server which receives the connections.
The TCPWRAP
is the new connection from the client. When a new\nconnection is made, the TCPWrap
instance is immediately constructed. This\nhappens outside of any JavaScript stack. (An executionAsyncId()
of 0
means\nthat it is being executed from C++ with no JavaScript stack above it.) With only\nthat information, it would be impossible to link resources together in\nterms of what caused them to be created, so triggerAsyncId
is given the task\nof propagating what resource is responsible for the new resource's existence.
resource
is an object that represents the actual async resource that has\nbeen initialized. This can contain useful information that can vary based on\nthe value of type
. For instance, for the GETADDRINFOREQWRAP
resource type,\nresource
provides the hostname used when looking up the IP address for the\nhost in net.Server.listen()
. The API for accessing this information is\ncurrently not considered public, but using the Embedder API, users can provide\nand document their own resource objects. For example, such a resource object\ncould contain the SQL query being executed.
In the case of Promises, the resource
object will have an\nisChainedPromise
property, set to true
if the promise has a parent promise,\nand false
otherwise. For example, in the case of b = a.then(handler)
, a
is\nconsidered a parent Promise
of b
. Here, b
is considered a chained promise.
In some cases the resource object is reused for performance reasons, it is\nthus not safe to use it as a key in a WeakMap
or add properties to it.
The following is an example with additional information about the calls to\ninit
between the before
and after
calls, specifically what the\ncallback to listen()
will look like. The output formatting is slightly more\nelaborate to make calling context easier to see.
let indent = 0;\nasync_hooks.createHook({\n init(asyncId, type, triggerAsyncId) {\n const eid = async_hooks.executionAsyncId();\n const indentStr = ' '.repeat(indent);\n fs.writeSync(\n 1,\n `${indentStr}${type}(${asyncId}):` +\n ` trigger: ${triggerAsyncId} execution: ${eid}\\n`);\n },\n before(asyncId) {\n const indentStr = ' '.repeat(indent);\n fs.writeFileSync('log.out',\n `${indentStr}before: ${asyncId}\\n`, { flag: 'a' });\n indent += 2;\n },\n after(asyncId) {\n indent -= 2;\n const indentStr = ' '.repeat(indent);\n fs.writeFileSync('log.out',\n `${indentStr}after: ${asyncId}\\n`, { flag: 'a' });\n },\n destroy(asyncId) {\n const indentStr = ' '.repeat(indent);\n fs.writeFileSync('log.out',\n `${indentStr}destroy: ${asyncId}\\n`, { flag: 'a' });\n },\n}).enable();\n\nrequire('net').createServer(() => {}).listen(8080, () => {\n // Let's wait 10ms before logging the server started.\n setTimeout(() => {\n console.log('>>>', async_hooks.executionAsyncId());\n }, 10);\n});\n
\nOutput from only starting the server:
\nTCPSERVERWRAP(5): trigger: 1 execution: 1\nTickObject(6): trigger: 5 execution: 1\nbefore: 6\n Timeout(7): trigger: 6 execution: 6\nafter: 6\ndestroy: 6\nbefore: 7\n>>> 7\n TickObject(8): trigger: 7 execution: 7\nafter: 7\nbefore: 8\nafter: 8\n
\nAs illustrated in the example, executionAsyncId()
and execution
each specify\nthe value of the current execution context; which is delineated by calls to\nbefore
and after
.
Only using execution
to graph resource allocation results in the following:
Timeout(7) -> TickObject(6) -> root(1)\n
\nThe TCPSERVERWRAP
is not part of this graph, even though it was the reason for\nconsole.log()
being called. This is because binding to a port without a\nhostname is a synchronous operation, but to maintain a completely asynchronous\nAPI the user's callback is placed in a process.nextTick()
.
The graph only shows when a resource was created, not why, so to track\nthe why use triggerAsyncId
.
asyncId
<number>When an asynchronous operation is initiated (such as a TCP server receiving a\nnew connection) or completes (such as writing data to disk) a callback is\ncalled to notify the user. The before
callback is called just before said\ncallback is executed. asyncId
is the unique identifier assigned to the\nresource about to execute the callback.
The before
callback will be called 0 to N times. The before
callback\nwill typically be called 0 times if the asynchronous operation was cancelled\nor, for example, if no connections are received by a TCP server. Persistent\nasynchronous resources like a TCP server will typically call the before
\ncallback multiple times, while other operations like fs.open()
will call\nit only once.
asyncId
<number>Called immediately after the callback specified in before
is completed.
If an uncaught exception occurs during execution of the callback, then after
\nwill run after the 'uncaughtException'
event is emitted or a domain
's\nhandler runs.
asyncId
<number>Called after the resource corresponding to asyncId
is destroyed. It is also\ncalled asynchronously from the embedder API emitDestroy()
.
Some resources depend on garbage collection for cleanup, so if a reference is\nmade to the resource
object passed to init
it is possible that destroy
\nwill never be called, causing a memory leak in the application. If the resource\ndoes not depend on garbage collection, then this will not be an issue.
asyncId
<number>Called when the resolve
function passed to the Promise
constructor is\ninvoked (either directly or through other means of resolving a promise).
resolve()
does not do any observable synchronous work.
The Promise
is not necessarily fulfilled or rejected at this point if the\nPromise
was resolved by assuming the state of another Promise
.
new Promise((resolve) => resolve(true)).then((a) => {});\n
\ncalls the following callbacks:
\ninit for PROMISE with id 5, trigger id: 1\n promise resolve 5 # corresponds to resolve(true)\ninit for PROMISE with id 6, trigger id: 5 # the Promise returned by then()\n before 6 # the then() callback is entered\n promise resolve 6 # the then() callback resolves the promise by returning\n after 6\n
",
"type": "module",
"displayName": "`promiseResolve(asyncId)`"
}
],
"type": "module",
"displayName": "Hook Callbacks"
},
{
"textRaw": "`async_hooks.executionAsyncId()`",
"name": "`async_hooks.executionasyncid()`",
"meta": {
"added": [
"v8.1.0"
],
"changes": [
{
"version": "v8.2.0",
"pr-url": "https://github.com/nodejs/node/pull/13490",
"description": "Renamed from `currentId`"
}
]
},
"desc": "asyncId
of the current execution context. Useful to\ntrack when something calls.const async_hooks = require('async_hooks');\n\nconsole.log(async_hooks.executionAsyncId()); // 1 - bootstrap\nfs.open(path, 'r', (err, fd) => {\n console.log(async_hooks.executionAsyncId()); // 6 - open()\n});\n
\nThe ID returned from executionAsyncId()
is related to execution timing, not\ncausality (which is covered by triggerAsyncId()
):
const server = net.createServer((conn) => {\n // Returns the ID of the server, not of the new connection, because the\n // callback runs in the execution scope of the server's MakeCallback().\n async_hooks.executionAsyncId();\n\n}).listen(port, () => {\n // Returns the ID of a TickObject (i.e. process.nextTick()) because all\n // callbacks passed to .listen() are wrapped in a nextTick().\n async_hooks.executionAsyncId();\n});\n
\nPromise contexts may not get precise executionAsyncIds
by default.\nSee the section on promise execution tracking.
const server = net.createServer((conn) => {\n // The resource that caused (or triggered) this callback to be called\n // was that of the new connection. Thus the return value of triggerAsyncId()\n // is the asyncId of \"conn\".\n async_hooks.triggerAsyncId();\n\n}).listen(port, () => {\n // Even though all callbacks passed to .listen() are wrapped in a nextTick()\n // the callback itself exists because the call to the server's .listen()\n // was made. So the return value would be the ID of the server.\n async_hooks.triggerAsyncId();\n});\n
\nPromise contexts may not get valid triggerAsyncId
s by default. See\nthe section on promise execution tracking.
By default, promise executions are not assigned asyncId
s due to the relatively\nexpensive nature of the promise introspection API provided by\nV8. This means that programs using promises or async
/await
will not get\ncorrect execution and trigger ids for promise callback contexts by default.
const ah = require('async_hooks');\nPromise.resolve(1729).then(() => {\n console.log(`eid ${ah.executionAsyncId()} tid ${ah.triggerAsyncId()}`);\n});\n// produces:\n// eid 1 tid 0\n
\nObserve that the then()
callback claims to have executed in the context of the\nouter scope even though there was an asynchronous hop involved. Also,\nthe triggerAsyncId
value is 0
, which means that we are missing context about\nthe resource that caused (triggered) the then()
callback to be executed.
Installing async hooks via async_hooks.createHook
enables promise execution\ntracking:
const ah = require('async_hooks');\nah.createHook({ init() {} }).enable(); // forces PromiseHooks to be enabled.\nPromise.resolve(1729).then(() => {\n console.log(`eid ${ah.executionAsyncId()} tid ${ah.triggerAsyncId()}`);\n});\n// produces:\n// eid 7 tid 6\n
\nIn this example, adding any actual hook function enabled the tracking of\npromises. There are two promises in the example above; the promise created by\nPromise.resolve()
and the promise returned by the call to then()
. In the\nexample above, the first promise got the asyncId
6
and the latter got\nasyncId
7
. During the execution of the then()
callback, we are executing\nin the context of promise with asyncId
7
. This promise was triggered by\nasync resource 6
.
Another subtlety with promises is that before
and after
callbacks are run\nonly on chained promises. That means promises not created by then()
/catch()
\nwill not have the before
and after
callbacks fired on them. For more details\nsee the details of the V8 PromiseHooks API.
Library developers that handle their own asynchronous resources performing tasks\nlike I/O, connection pooling, or managing callback queues may use the\nAsyncWrap
JavaScript API so that all the appropriate callbacks are called.
The class AsyncResource
is designed to be extended by the embedder's async\nresources. Using this, users can easily trigger the lifetime events of their\nown resources.
The init
hook will trigger when an AsyncResource
is instantiated.
The following is an overview of the AsyncResource
API.
const { AsyncResource, executionAsyncId } = require('async_hooks');\n\n// AsyncResource() is meant to be extended. Instantiating a\n// new AsyncResource() also triggers init. If triggerAsyncId is omitted then\n// async_hook.executionAsyncId() is used.\nconst asyncResource = new AsyncResource(\n type, { triggerAsyncId: executionAsyncId(), requireManualDestroy: false }\n);\n\n// Run a function in the execution context of the resource. This will\n// * establish the context of the resource\n// * trigger the AsyncHooks before callbacks\n// * call the provided function `fn` with the supplied arguments\n// * trigger the AsyncHooks after callbacks\n// * restore the original execution context\nasyncResource.runInAsyncScope(fn, thisArg, ...args);\n\n// Call AsyncHooks destroy callbacks.\nasyncResource.emitDestroy();\n\n// Return the unique ID assigned to the AsyncResource instance.\nasyncResource.asyncId();\n\n// Return the trigger ID for the AsyncResource instance.\nasyncResource.triggerAsyncId();\n
",
"modules": [
{
"textRaw": "`new AsyncResource(type[, options])`",
"name": "`new_asyncresource(type[,_options])`",
"desc": "type
<string> The type of async event.options
<Object>
triggerAsyncId
<number> The ID of the execution context that created this\nasync event. Default: executionAsyncId()
.requireManualDestroy
<boolean> Disables automatic emitDestroy
when the\nobject is garbage collected. This usually does not need to be set (even if\nemitDestroy
is called manually), unless the resource's asyncId
is\nretrieved and the sensitive API's emitDestroy
is called with it.\nDefault: false
.Example usage:
\nclass DBQuery extends AsyncResource {\n constructor(db) {\n super('DBQuery');\n this.db = db;\n }\n\n getInfo(query, callback) {\n this.db.get(query, (err, data) => {\n this.runInAsyncScope(callback, null, err, data);\n });\n }\n\n close() {\n this.db = null;\n this.emitDestroy();\n }\n}\n
",
"type": "module",
"displayName": "`new AsyncResource(type[, options])`"
},
{
"textRaw": "`asyncResource.runInAsyncScope(fn[, thisArg, ...args])`",
"name": "`asyncresource.runinasyncscope(fn[,_thisarg,_...args])`",
"meta": {
"added": [
"v9.6.0"
],
"changes": []
},
"desc": "fn
<Function> The function to call in the execution context of this async\nresource.thisArg
<any> The receiver to be used for the function call....args
<any> Optional arguments to pass to the function.Call the provided function with the provided arguments in the execution context\nof the async resource. This will establish the context, trigger the AsyncHooks\nbefore callbacks, call the function, trigger the AsyncHooks after callbacks, and\nthen restore the original execution context.
", "type": "module", "displayName": "`asyncResource.runInAsyncScope(fn[, thisArg, ...args])`" }, { "textRaw": "`asyncResource.emitDestroy()`", "name": "`asyncresource.emitdestroy()`", "desc": "asyncResource
.Call all destroy
hooks. This should only ever be called once. An error will\nbe thrown if it is called more than once. This must be manually called. If\nthe resource is left to be collected by the GC then the destroy
hooks will\nnever be called.
asyncId
assigned to the resource.triggerAsyncId
that is passed to the\nAsyncResource
constructor.