A brute-force protection middleware for express routes that rate-limits incoming requests, increasing the delay with each request in a fibonacci-like sequence.
via npm:
$ npm install express-brute
var ExpressBrute = require('express-brute'),
var store = new ExpressBrute.MemoryStore(); // stores state locally, don't use this in production
var bruteforce = new ExpressBrute(store);
app.post('/auth',
bruteforce.prevent, // error 429 if we hit this route too often
function (req, res, next) {
res.send('Success!');
}
);
store
An instance ofExpressBrute.MemoryStore
or some other ExpressBrute store (see a list of known stores below).options
freeRetries
The number of retires the user has before they need to start waiting (default: 2)minWait
The initial wait time (in milliseconds) after the user runs out of retries (default: 500 milliseconds)maxWait
The maximum amount of time (in milliseconds) between requests the user needs to wait (default: 15 minutes). The wait for a given request is determined by adding the time the user needed to wait for the previous two requests.lifetime
The length of time (in seconds since the last request) to remember the number of requests that have been made by an IP. By default it will be set tomaxWait * the number of attempts before you hit maxWait
to discourage simply waiting for the lifetime to expire before resuming an attack. With default values this is about 6 hours.failCallback
Gets called with (req
,resp
,next
,nextValidRequestDate
) when a request is rejected (default: ExpressBrute.FailForbidden)proxyDepth
Specifies how many levels of theX-Forwarded-For
header to trust. If your web server is behind a CDN and/or load balancer you'll need to set this to however many levels of proxying it's behind to get a valid IP. Setting this too high allows attackers to get around brute force protection by spoofing theX-Forwarded-For
header, so don't set it higher than you need to (default: 0)attachResetToRequest
Specify whether or not a simplified reset method should be attached atreq.brute.reset
. The simplified method takes only a callback, and resets allExpressBrute
middleware that was called on the current request. If multiple instances ofExpressBrute
have middleware on the same request, only those withattachResetToRequest
set to true will be reset (default: true)refreshTimeoutOnRequest
Defines whether the remaininglifetime
of a counter should be based on the time since the last request (true) of the time since the first request (false). Useful for allowing limits over fixed periods of time, for example a limited number of requests per day. (Default: true)handleStoreError
Gets called whenever an error occurs with the persistent store from which ExpressBrute cannot recover. It is passed an object containing the propertiesmessage
(a description of the message),parent
(the error raised by the session store), and [key
,ip
] or [req
,res
,next
] depending on whether or the error occurs duringreset
or in the middleware itself.
An in-memory store for persisting request counts. Don't use this in production, instead choose one of the more robust store implementations listed below.
prevent(req, res, next)
Middleware that will bounce requests that happen faster than the current wait time by callingfailCallback
. Equivilent togetMiddleware(null)
getMiddleware(options)
Generates middleware that will bounce requests with the samekey
and IP address that happen faster than the current wait time by callingfailCallback
. Also attaches a function atreq.brute.reset
that can be called to reset the counter for the current ip and key. This functions the thereset
instance method, but without the need to explicitly pass theip
andkey
paramterskey
can be a string or alternatively it can be afunction(req, res, next)
that or callsnext
, passing a string as the first parameter.failCallback
Allows you to override the value offailCallback
for this middleware
reset(ip, key, next)
Resets the wait time between requests back to its initial value. You can passnull
forkey
if you want to reset a request protected byprotect
.getIPFromRequest(req)
Uses the current proxy trust settings to get the current IP from a request object
There are some built-in callbacks that come with BruteExpress that handle some common use cases.
ExpressBrute.FailTooManyRquests
Terminates the request and responses with a 429 (Too Many Requests) error that has aRetry-After
header and a JSON error message.ExpressBrute.FailForbidden
Terminates the request and responds with a 403 (Forbidden) error that has aRetry-After
header and a JSON error message. This is provided for compatibility with ExpressBrute versions prior to v0.5.0, for new usersFailTooManyRequests
is the preferred behavior.ExpressBrute.FailMark
Sets res.nextValidRequestDate, the Retry-After header and the res.status=429, then calls next() to pass the request on to the appropriate routes.
There are a number adapters that have been written to allow ExpressBrute to be used with different persistent storage implementations, some of the ones I know about include:
If you write your own store and want me to add it to the list, just drop me an email or create an issue.
require('connect-flash');
var ExpressBrute = require('express-brute'),
MemcachedStore = require('express-brute-memcached'),
moment = require('moment'),
store;
if (config.environment == 'development'){
store = new ExpressBrute.MemoryStore(); // stores state locally, don't use this in production
} else {
// stores state with memcached
store = new MemcachedStore(['127.0.0.1'], {
prefix: 'NoConflicts'
});
}
var failCallback = function (req, res, next, nextValidRequestDate) {
req.flash('error', "You've made too many failed attempts in a short period of time, please try again "+moment(nextValidRequestDate).fromNow());
res.redirect('/login'); // brute force protection triggered, send them back to the login page
};
var handleStoreError = handleStoreError: function (error) {
log.error(error); // log this error so we can figure out what went wrong
// cause node to exit, hopefully restarting the process fixes the problem
throw {
message: error.message,
parent: error.parent
};
}
// Start slowing requests after 5 failed attempts to do something for the same user
var userBruteforce = new ExpressBrute(store, {
freeRetries: 5,
proxyDepth: 1,
minWait: 5*60*1000, // 5 minutes
maxWait: 60*60*1000, // 1 hour,
failCallback: failCallback,
handleStoreError: handleStoreError
}
});
// No more than 1000 login attempts per day per IP
var globalBruteforce = new ExpressBrute(store, {
freeRetries: 1000,
proxyDepth: 1,
attachResetToRequest: false,
refreshTimeoutOnRequest: false,
minWait: 25*60*60*1000, // 1 day 1 hour (should never reach this wait time)
maxWait: 25*60*60*1000, // 1 day 1 hour (should never reach this wait time)
lifetime: 24*60*60, // 1 day (seconds not milliseconds)
failCallback: failCallback,
handleStoreError: handleStoreError
});
app.post('/auth',
globalBruteforce.prevent,
userBruteforce.getMiddleware({
key: function(req, res, next) {
// prevent too many attempts for the same username
next(req.body.username);
}
}),
function (req, res, next) {
if (User.isValidLogin(req.body.username, req.body.password)) { // omitted for the sake of conciseness
// reset the failure counter so next time they log in they get 5 tries again before the delays kick in
req.brute.reset(function () {
res.redirect('/'); // logged in, send them to the home page
});
} else {
res.flash('error', "Invalid username or password")
res.redirect('/login'); // bad username/password, send them back to the login page
}
}
);
- NEW: Added the
handleStoreError
option to allow more customizable handling of errors that are thrown by the persistent store. Default behavior is to throw the errors as an exception - there is nothing ExpressBrute can do to recover. - CHANGED: Errors thrown as a result of errors raised by the store now include the store's error as well, for debugging purposes.
- CHANGED: Stopped using res.send(status, body), as it is deprecated in express 4.x. Instead call res.status and res.send separately (Thanks marinewater!)
- BUG: When setting proxyDepth to 1, ips is never populated with proxied X-Forwarded-For IP.
- NEW: Added an additional
FailTooManyRequests
failure callback, that returns a 429 (TooManyRequests) error instead of 403 (Forbidden). This is a more accurate error status code. - NEW: All the built in failure callbacks now set the "Retry-After" header to the number of seconds until it is safe to try again. Per RFC6585
- NEW: Documentation updated to list some known store implementations.
- CHANGED: Default failure callback is now
FailTooManyRequests
.FailForbidden
remains an option for backwards compatiblity. - CHANGED: ExpressBrute.MemcachedStore is no longer included by default, and is now available as a separate module (because there are multiple store options it doesn't really make sense to include one by default).
- CHANGED:
FailMark
no longer sets returns 403 Forbidden, instead does 429 TooManyRequets.
- BUG: In some cases when no callbacks were supplied memcached would drop the request. Ensure that memcached always sees a callback even if ExpressBrute isn't given one.
- NEW:
refreshTimeoutOnRequest
option that allows you to prevent the remaininglifetime
for a timer from being reset on each request (useful for implementing limits for set time frames, e.g. requests per day) - BUG: Lifetimes were not previously getting extended properly for instances of
ExpressBrute.MemoryStore
- NEW:
attachResetToRequest
parameter that lets you prevent the request object being decorated - NEW:
failCallback
can be overriden bygetMiddleware
- NEW:
proxyDepth
option onExpressBrute
that specifies how many levels of theX-Forwarded-For
header to trust (inspired by express-bouncer). - NEW:
getIPFromRequest
method that essentially allowsreset
to used in a similar ways as in v0.2.2. This also respects the newproxyDepth
setting. - CHANGED:
getMiddleware
now takes an options object instead of the key directly.
- NEW: Support for using custom keys to group requests further (e.g. grouping login requests by username)
- NEW: Support for middleware from multiple instances of
ExpressBrute
on the same route. - NEW: Tracking
lifetime
now has a reasonable default derived from the other settings for that instance ofExpressBrute
- NEW: Keys are now hashed before saving to a store, to prevent really long key names and reduce the possibility of collisions.
- NEW: There is now a convience method that gets attached to
req
object asreq.brute.reset
. It takes a single parameter (a callback), and will reset all the counters used byExpressBrute
middleware that was called for the current route. - CHANGED: Tracking
lifetime
is now specified onExpressBrute
instead ofMemcachedStore
. This also means lifetime is now supported by MemoryStore. - CHANGED: The function signature for
ExpressBrute.reset
has changed. It now requires an IP and key be passed instead of a request object. - IMPROVED: Efficiency for large values of
freeRetries
. - BUG: Removed a small chance of incorrectly triggering brute force protection.