bottleneck

Bottleneck is a lightweight and efficient Task Scheduler and Rate Limiter for Node.js and the browser. When dealing with services with limited resources, it's important to ensure that they don't become overloaded.

Bottleneck is an easy solution as it does not add much complexity to your code.

It is battle-hardened, reliable and production-ready. Hundreds of projects rely on it and it is used on a large scale in private companies and open source software.

It also supports distributed applications through the new Clustering feature in v2.

Bottleneck Version 2

This new version is almost 100% compatible with v1 and adds powerful features such as:

True Clustering support. You can now rate limit and schedule jobs across multiple Node.js instances. It uses strictly atomic operations to stay reliable in the presence of unreliable clients. 100% of Bottleneck's features are supported.
Support for custom job weights. Not all jobs are equally resource intensive.
Support for job expirations. Bottleneck can automatically cancel jobs if they exceed their execution time limit.
Many improvements to the interface, such as better method names and errors, improved debugging tools.

Quickly upgrade your application code from v1 to v2 of Bottleneck

Bottleneck v2 targets Node v6.0 or newer, and evergreen browsers. Using an older version of Node? You should upgrade very soon!

Bottleneck v1 targets ES5, which makes it compatible with any browser or Node version. It's still maintained, but it will not be receiving any new features. Browse the v1 documentation.

Install

npm install --save bottleneck

Quick Start

Most APIs have a rate limit. For example, the reddit.com API limits scripts to 1 request every 2 seconds.

const Bottleneck = require("bottleneck");

// Never more than 1 request running at a time.
// Wait at least 2000ms between each request.
const limiter = new Bottleneck({
  maxConcurrent: 1,
  minTime: 2000
});

Instead of this:

someAsyncCall(arg1, arg2, callback);

Do this:

limiter.submit(someAsyncCall, arg1, arg2, callback);

And now you can be assured that someAsyncCall will abide by your rate guidelines!

More information about using Bottleneck with callbacks

Promises example

Instead of this:

myFunction(arg1, arg2)
.then((result) => { /* handle result */ })

Do this:

limiter.schedule(() => myFunction(arg1, arg2))
.then((result) => { /* handle result */ })

Or this:

const wrapped = limiter.wrap(myFunction)

wrapped(arg1, arg2)
.then((result) => { /* handle result */ })

More information about using Bottleneck with promises

Remember...

Bottleneck builds a queue of jobs and executes them as soon as possible. All the jobs will be executed in the order that they were received. See priorities if you wish to alter this behavior.

This is sufficient for the vast majority of applications. Read the 'Gotchas' section and you're good to go. Or keep reading to learn about all the fine tuning available for the more complex use cases.

Gotchas

Make sure you're catching error events emitted by limiters! See Debugging your application
When using submit, if a callback isn't necessary, you must pass null or an empty function instead. It will not work otherwise.
When using submit, make sure all the jobs will eventually complete by calling their callback (or have an expiration). Again, even if you submitted your job with a null callback , it still needs to call its callback. This is particularly important if you are using a maxConcurrent value that isn't null (unlimited), otherwise those uncompleted jobs will be clogging up the limiter and no new jobs will be able to run. It's safe to call the callback more than once, subsequent calls are ignored.
When using schedule or wrap, make sure that all the jobs will eventually complete (resolving or rejecting) or have an expiration. This is very important if you are using a maxConcurrent value that isn't null (unlimited), otherwise those uncompleted jobs will be clogging up the limiter and no new jobs will be able to run.
Clustering has its own share of gotchas. Read the Clustering chapter carefully.

Docs

Constructor

const limiter = new Bottleneck({ /* options */ });

Basic options:

Option	Default	Description
`maxConcurrent`	`null` (unlimited)	How many jobs can be running at the same time.
`minTime`	`0` (ms)	How long to wait after launching a job before launching another one.
`highWater`	`null`	How long can the queue get? When the queue length exceeds that value, the selected `strategy` is executed to shed the load.
`strategy`	`Bottleneck.strategy.LEAK`	Which strategy to use if the queue gets longer than the high water mark. Read about strategies.
`penalty`	`15 * minTime`, or `5000` when `minTime` is `null`	The `penalty` value used by the `Bottleneck.strategy.BLOCK` strategy.
`reservoir`	`null` (unlimited)	How many jobs can be executed before the limiter stops executing jobs. If `reservoir` reaches `0`, no jobs will be executed until it is no longer `0`.

submit()

Adds a job to the queue. This is the callback version of schedule.

limiter.submit(someAsyncCall, arg1, arg2, callback);

submit can also accept some advanced options. See Job Options.

It's safe to mix submit and schedule in the same limiter.

schedule()

Adds a job to the queue. This is the Promise version of submit.

const fn = function(arg1, arg2) {
  return httpGet(arg1, arg2); // Here httpGet() returns a promise
};

limiter.schedule(fn, arg1, arg2)
.then((result) => {
  /* ... */
})

Simply put, schedule takes a function Fn and a list of arguments. Fn must return a promise. schedule returns a promise that will be executed according to the rate limits.

schedule can also accept some advanced options. See Job Options.

It's safe to mix submit and schedule in the same limiter.

Here's another example:

// suppose that `http.get(url)` returns a promise

const url = "https://wikipedia.org";

limiter.schedule(() => http.get(url))
.then(response => console.log(response.body));

If your function does not return a promise, it needs to use Promise.resolve like so:

// GOOD
limiter.schedule(() => Promise.resolve("This is a string"))
.then(data => console.log(data));

// INCORRECT!
limiter.schedule(() => "This is a string")
.then(data => console.log(data));

It's also possible to replace the Promise library used by Bottleneck:

const Bottleneck = require("bottleneck");
const Promise = require("bluebird");

const limiter = new Bottleneck({
  /* other options... */
  Promise: Promise
});

wrap()

Takes a function that returns a promise. Returns a function identical to the original, but rate limited.

const wrapped = limiter.wrap(fn)

wrapped()
.then(function (result) {
  /* ... */
})
.catch(function (error) {
  // Bottleneck might need to fail the job even if the original function can never fail
})

Job Options

Both submit and schedule accept advanced options.

// Submit
limiter.submit(options, someAsyncCall, arg1, arg2, callback);

// Schedule
limiter.schedule(options, fn, arg1, arg2);

Option	Default	Description
`priority`	`5`	A priority between `0` and `9`. A job with a priority of `4` will always be executed before a job with a priority of `5`.
`weight`	`1`	Must be an integer equal to or higher than `0`. The `weight` is what increases the number of running jobs (up to `maxConcurrent`, if using) and decreases the `reservoir` value (if using).
`expiration`	`null` (unlimited)	The number milliseconds a job has to finish. Jobs that take longer than their `expiration` will be failed with a `BottleneckError`.
`id`	`<no-id>`	You can give an ID to your jobs, for easier debugging. See Debugging your application.

Strategies

A strategy is a simple algorithm that is executed every time adding a job would cause the number of queued jobs to exceed highWater. See Events.

Bottleneck.strategy.LEAK

When adding a new job to a limiter, if the queue length reaches highWater, drop the oldest job with the lowest priority. This is useful when jobs that have been waiting for too long are not important anymore. If all the queued jobs are more important (based on their priority value) than the one being added, it will not be added.

Bottleneck.strategy.OVERFLOW_PRIORITY

Same as LEAK, except it will only drop jobs that are less important than the one being added. If all the queued jobs are as or more important than the new one, it will not be added.

Bottleneck.strategy.OVERFLOW

When adding a new job to a limiter, if the queue length reaches highWater, do not add the new job. This strategy totally ignores priority levels.

Bottleneck.strategy.BLOCK

When adding a new job to a limiter, if the queue length reaches highWater, the limiter falls into "blocked mode". All queued jobs are dropped and no new jobs will be accepted until the limiter unblocks. It will unblock after penalty milliseconds have passed without receiving a new job. penalty is equal to 15 * minTime (or 5000 if minTime is 0) by default and can be changed by calling changePenalty(). This strategy is ideal when bruteforce attacks are to be expected. This strategy totally ignores priority levels.

Jobs lifecycle

Received. You new job has been added to your limiter. Bottleneck needs to check whether it can be accepted into the queue, based on your highWater setting.
Queued. Bottleneck has accepted your job, but it can not tell at what exact timestamp it will run yet, because it is dependent on previous jobs.
Running. Your job is not in the queue anymore, it will be executed after a delay computed according to your minTime setting.
Executing. Your job is executing its code.
Done. Your job has completed.

counts()

const counts = limiter.counts();

console.log(counts);
/*
{
  RECEIVED: 0,
  QUEUED: 0,
  RUNNING: 0,
  EXECUTING: 0,
  DONE: 0
}
*/

Returns an object with the current number of jobs per status in the limiter.

Note: By default, Bottleneck does not keep track of DONE jobs, to save memory. You can enable that feature by passing trackDoneStatus: true as an option when creating a limiter.

jobStatus()

console.log(limiter.jobStatus("some-job-id"));
// Example: QUEUED

Returns the status of the job with the provided job id. See Job Options.

Note: By default, Bottleneck does not keep track of DONE jobs, to save memory. You can enable that feature by passing trackDoneStatus: true as an option when creating a limiter.

queued()

const count = limiter.queued(priority);

console.log(count);

priority is optional. Returns the number of Queued jobs with the given priority level. Omitting the priority argument returns the total number of queued jobs in the limiter.

empty()

if (limiter.empty()) {
  // do something...
}

Returns a boolean which indicates whether there are any Received or Queued jobs in the limiter.

running()

limiter.running()
.then((count) => console.log(count));

Returns a promise that returns the total weight of the Running and Executing jobs in the Cluster.

check()

limiter.check()
.then((wouldRunNow) => console.log(wouldRunNow));

Checks if a new job would be executed immediately if it was submitted now. Returns a promise that returns a boolean.

Events

Event names: error, empty, idle, dropped and debug.

error

limiter.on('error', function (error) {
  // This is where Bottleneck's errors end up.
})

By far the most common case for errors is uncaught exceptions in your application code. If the jobs you add to Bottleneck (through submit, schedule, wrap, etc.) don't catch their own exceptions, the limiter will emit an error event.

If using Clustering, errors thrown by the Redis client will emit an error event.

empty

limiter.on('empty', function () {
  // This will be called when `limiter.empty()` becomes true.
})

idle

limiter.on('idle', function () {
  // This will be called when `limiter.empty()` is `true` and `limiter.running()` is `0`.
})

dropped

limiter.on('dropped', function (dropped) {
  // This will be called when a strategy was triggered.
  // The dropped request is passed to this event listener.
})

depleted

limiter.on('depleted', function (empty) {
  // This will be called every time the reservoir drops to 0.
  // The `empty` (boolean) argument indicates whether `limiter.empty()` is currently true.
})

debug

limiter.on('debug', function (message, data) {
  // Useful to figure out what the limiter is doing in real time
  // and to help debug your application
})

Use removeAllListeners() with an optional event name as first argument to remove listeners.

Use .once() instead of .on() to only receive a single event.

updateSettings()

limiter.updateSettings(options);

The options are the same as the limiter constructor.

Note: This doesn't affect jobs already scheduled for execution.

For example, imagine that three 60-second jobs are added to a limiter at time T with maxConcurrent: null and minTime: 2000. The jobs will be launched at T seconds, T+2 seconds and T+4 seconds, respectively. If right after adding the jobs to Bottleneck, you were to call limiter.updateSettings({ maxConcurrent: 1 });, it won't change the fact that there will be 3 jobs running at the same time for roughly 60 seconds. updateSettings() only affects jobs that have not yet been added.

This is by design, as Bottleneck made a promise to execute those requests according to the settings valid at the time.

incrementReservoir()

limiter.incrementReservoir(incrementBy);

This is a way to update the reservoir value other than calling updateSettings.

Returns a promise.

currentReservoir()

limiter.currentReservoir()
.then((reservoir) => console.log(reservoir));

Returns a promise that returns the current reservoir value.

chain()

limiter : If another limiter is passed, tasks that are ready to be executed will be added to that other limiter. Default: null (none)

Suppose you have 2 types of tasks, A and B. They both have their own limiter with their own settings, but both must also follow a global limiter C:

var limiterA = new Bottleneck( /* ...some settings... */ );
var limiterB = new Bottleneck( /* ...some different settings... */ );
var limiterC = new Bottleneck( /* ...some global settings... */ );
limiterA.chain(limiterC);
limiterB.chain(limiterC);
// Requests added to limiterA must follow the A and C rate limits.
// Requests added to limiterB must follow the B and C rate limits.
// Requests added to limiterC must follow the C rate limits.

To unchain, call limiter.chain(null);.

Clustering

Clustering lets many limiters access the same shared state, stored in a Redis server or Redis cluster. Changes to the state are Atomic, Consistent and Isolated (and fully ACID with the right redis Durability configuration), to eliminate any chances of race conditions or state corruption. Your settings, such as maxConcurrent, minTime, etc., are shared across the whole cluster, which means—for example—that { maxConcurrent: 5 } guarantees no more than 5 jobs can ever run at a time in the entire cluster of limiters. 100% of Bottleneck's features are supported in Clustering mode. Enabling Clustering is as simple as changing a few settings. It's also a convenient way to store or export state for later use.

Enabling Clustering

IMPORTANT: Add redis to your application's dependencies.

npm install --save redis

const limiter = new Bottleneck({
  /* Some basic options */
  maxConcurrent: 5,
  minTime: 500
  id: "my-super-app" // Should be unique for every limiter in the same Redis db

  /* Clustering options */
  datastore: "redis",
  clearDatastore: false,
  clientOptions: {
    // node-redis client options, passed to redis.createClient()
    // See https://github.com/NodeRedis/node_redis#options-object-properties
    host: "127.0.0.1",
    port: 6379
    // "db" is another useful option
  }
})

Option	Default	Description
`datastore`	`"local"`	Where the limiter stores its internal state. The default (`local`) keeps the state in the limiter itself. Set it to `redis` to enable Clustering.
`clearDatastore`	`false`	When set to `true`, on initial startup, the limiter will wipe any existing Bottleneck state data on the Redis db.
`clientOptions`	`{}`	This object is passed directly to NodeRedis's `redis.createClient()` method. See all the valid client options.

`.ready()`

Since your limiter has to connect to Redis, and this operation takes time and can fail, you need to wait for your limiter to be connected before using it.

const limiter = new Bottleneck({ /* ... */ })

limiter.ready()
.then(() => {
  // The limiter is ready
})
.catch((error) => {
  // The limiter couldn't start
})

The .ready() method also exists when using the local datastore, for code compatibility reasons: code written for redis will always work with local.

`.disconnect(flush)`

This helper method calls the .end(flush) method on the Redis clients used by a limiter.

limiter.disconnect(true)

The flush argument is optional and defaults to true.

`.clients()`

If you need direct access to the redis clients, use .clients():

console.log(limiter.clients())
// { client: <Redis Client>, subscriber: <Redis Client> }

Just like .ready(), calling .clients() when using the local datastore won't fail, it just won't return anything.

Important considerations when Clustering

The first limiter connecting to Redis will store its constructor options (Constructor) on Redis and all subsequent limiters will be using those settings. You can alter the constructor options used by all the connected limiters by calling updateSettings. The clearDatastore option instructs a new limiter to wipe any previous Bottleneck data, including previously stored settings.

Queued jobs are NOT stored on Redis. They are local to each limiter. Exiting the Node.js process will lose those jobs. This is because Bottleneck has no way to propagate the JS code to run a job across a different Node.js process than the one it originated on. Bottleneck doesn't keep track of the queue contents of the limiters on a cluster for performance and reliability reasons.

Due to the above, functionality relying on the queue length happen purely locally:

Priorities are local. A higher priority job will run before a lower priority job on the same limiter. Another limiter on the cluster might run a lower priority before our higher priority one.
(Assuming default priority levels) Bottleneck guarantees that jobs will be run in the order they were queued on the same limiter. Another limiter on the cluster might run a job queued later before ours runs.
highWater and load shedding (strategies) are per limiter. However, one limiter entering Blocked mode will put the entire cluster in Blocked mode until penalty milliseconds have passed. See Strategies.
The empty event is triggered when the (local) queue is empty.
The idle event is triggered when the (local) queue is empty and no jobs are currently running anywhere in the cluster.

You must work around these limitations in your application code if they are an issue to you.

The current design guarantees reliability and lets clients (limiters) come and go. Your application can scale up or down, and clients can be disconnected without issues.

It is strongly recommended that you give an id for every limiter since it is used to build the name of your limiter's Redis keys! Limiters with the same id inside the same Redis db will be sharing the same datastore!

It is strongly recommended that you set an expiration (See Job Options) on every job, since that lets the cluster recover from crashed or disconnected clients. Otherwise, a client crashing while executing a job would not be able to tell the cluster to decrease its number of "running" jobs. By using expirations, those lost jobs are automatically cleared after the specified time has passed. Using expirations is essential to keeping a cluster reliable in the face of unpredictable application bugs, network hiccups, and so on.

Network latency between Node.js and Redis is not taken into account when calculating timings (such as minTime). To minimize the impact of latency, Bottleneck performs the absolute minimum number of state accesses. Keeping the Redis server close to your limiters will help you get a more consistent experience. Keeping the clients' server time consistent will also help.

It is strongly recommended to set up an error listener.

Bottleneck does not guarantee that the concurrency will be spread evenly across limiters. With { maxConcurrent: 5 }, it's absolutely possible for a single limiter to end up running 5 jobs simultaneously while the other limiters in the cluster sit idle. To spread the load, use the .chain() method:

const clusterLimiter = new Bottleneck({ maxConcurrent: 5, datastore: 'redis' });
const limiter = new Bottleneck({ maxConcurrent: 1 });

limiter.chain(clusterLimiter);

clusterLimiter.ready()
.then(() => {
  // Any Node process can only run one job at a time.
  // Across the whole cluster, up to 5 jobs can run simultaneously.

  limiter.schedule( /* ... */ )
})
.catch((error) => { /* ... */ });

Additional Clustering information

At the moment, each limiter opens 2 connections to Redis. This can lead to a high number of connections, especially when Groups are used. This might change in a future release.
Bottleneck is compatible with Redis Clusters.
Bottleneck's data is stored in Redis keys starting with b_. It also uses pub/sub channels starting with b_ It will not interfere with any other data stored on the server.
Bottleneck loads a few Lua scripts on the Redis server using the SCRIPT LOAD command. These scripts only take up a few Kb of memory. Running the SCRIPT FLUSH command will cause any connected limiters to experience critical errors until a new limiter connects to Redis and loads the scripts again.
The Lua scripts are highly optimized and designed to use as few resources (CPU, especially) as possible.

Groups and Clustering

If you are using a Group, the generated limiters automatically receive an id with the pattern group-key-${KEY}.
A Group collects its own garbage, and so when using Clustering, it manages the Redis TTL (TTL) on the keys it uses to ensure they get cleaned up by Redis when unused for longer than the Group's timeout setting.
Each limiter opens 2 connections to Redis. Be careful not to go over Redis' maxclients value.

Debugging your application

Debugging complex scheduling logic can be difficult, especially when priorities, weights, and network latency all interact.

If your application is not behaving as expected, start by making sure you're catching error events emitted by your limiters. Those errors are most likely uncaught exceptions from your application code.

To see exactly what a limiter is doing in real time, listen to the debug event. It contains detailed information about how the limiter is executing your code. Adding job IDs to all your jobs makes the debug output more readable.

When Bottleneck has to fail one of your jobs, it does so by using BottleneckError objects. This lets you tell those errors apart from your own code's errors:

limiter.schedule(fn)
.then((result) => { /* ... */ } )
.catch((error) => {
  if (error instanceof Bottleneck.prototype.BottleneckError) {
    /* ... */
  }
})

Some Promise libraries also support selective catch() blocks that only catch a specific type of error:

limiter.schedule(fn)
.then((result) => { /* ... */ } )
.catch(Bottleneck.prototype.BottleneckError, (error) => {
  /* ... */
})
.catch((error) => {
  /* ... */
})

You can also set the constructor option rejectOnDrop to false, and Bottleneck will leave your failed jobs hanging instead of failing them.

Group

The Group feature of Bottleneck manages many limiters automatically for you. It creates limiters dynamically and transparently.

Let's take a DNS server as an example of how Bottleneck can be used. It's a service that sees a lot of abuse and where incoming DNS requests need to be rate limited. Bottleneck is so tiny, it's acceptable to create one limiter for each origin IP, even if it means creating thousands of limiters. The Group feature is perfect for this use case. Create one group and use the origin IP to rate limit each IP independently. Each call with the same key (IP) will be routed to the same underlying limiter. A group is created like a limiter:

const group = new Bottleneck.Group(options);

The options object will be used for every limiter created by the group.

The group is then used with the .key(str) method:

// In this example, the key is an IP
group.key("77.66.54.32").submit(someAsyncCall, arg1, arg2, cb);

key()

str : The key to use. All jobs added with the same key will use the same underlying limiter. Default: ""

The return value of .key(str) is a limiter. If it doesn't already exist, it is generated for you. Calling key() is how limiters are created inside a Group. Limiters that have been idle for longer than 5 minutes are deleted to avoid memory leaks.

on('created')

group.on('created', (limiter, key) => {
  console.log('A new limiter was created for key: ' + key)

  // Prepare the limiter, for example we'll want to listen to its 'error' events!
  limiter.on('error', (err) => {
    // Handle errors here
  })

  //
  // ...other operations to be executed when a new limiter is created...
  //
})

Listening for the created event is the recommended way to set up a new limiter. Your event handler is executed before key() returns the newly created limiter.

updateSettings()

group.updateSettings({ timeout: 60000 })

timeout: The expiration time for unused limiters, in milliseconds. By default, it is 300000 (5 minutes).

When autocleanup is enabled, limiters not used in the last timeout milliseconds will be deleted to avoid memory leaks.

deleteKey()

str: The key for the limiter to delete.

Manually deletes the limiter at the specified key. This can be useful when the auto cleanup is turned off.

keys()

Returns an array containing all the keys in the group.

limiters()

const limiters = group.limiters()

console.log(limiters)
// [ { key: "some key", limiter: <limiter> }, { key: "some other key", limiter: <some other limiter> } ]

Upgrading to v2

The internal algorithms essentially haven't changed from v1, but many small changes to the interface were made to introduce new features.

All the breaking changes:

Bottleneck v2 uses ES6/ES2015. v1 will continue to use ES5 only.
The Bottleneck constructor now takes an options object. See Constructor.
Jobs take an optional options object. See Job options.
Removed submitPriority(), use submit() with an options object instead.
Removed schedulePriority(), use schedule() with an options object instead.
The rejectOnDrop option is now true by default.
Use null instead of 0 to indicate an unlimited maxConcurrent value.
Use null instead of -1 to indicate an unlimited highWater value.
Renamed changeSettings() to updateSettings(), it now returns a promise to indicate completion. It takes the same options object as the constructor.
Renamed nbQueued() to queued().
Renamed nbRunning to running(), it now returns its result using a promise.
Removed isBlocked().
Changing the Promise library is now done through the options object like any other limiter setting.
Removed changePenalty(), it is now done through the options object like any other limiter setting.
Removed changeReservoir(), it is now done through the options object like any other limiter setting.
Removed stopAll(). Use the reservoir feature to disable execution instead.
check() now accepts an optional weight argument, and returns its result using a promise.
The Cluster feature is now called Group. This is to distinguish it from the new v2 Clustering feature.
The Group constructor takes an options object to match the limiter constructor.
Renamed the Group changeTimeout() method to updateSettings(), it now takes an options object. See Group.

Version 2 is more user-friendly, powerful and reliable.

After upgrading your code, please take a minute to read the Debugging your application chapter.

Contributing

This README is always in need of improvements. If wording can be clearer and simpler, please consider forking this repo and submitting a Pull Request, or simply opening an issue.

Suggestions and bug reports are also welcome.

To work on the Bottleneck code, simply clone the repo, and run ./scripts/build.sh && npm test to ensure that everything is set up correctly.

Make your changes to the files located in src/ only, then run ./scripts/build.sh && npm test to build and test them.

To speed up compilation time during development, run ./scripts/build.sh dev instead. Make sure to build and test without dev before submitting a PR.

The tests must also pass in Clustering mode. You'll need a Redis server running on 127.0.0.1:6379, then run ./scripts/build.sh && DATASTORE=redis npm test.

All contributions are appreciated and will be considered.