Replays a request when a network error occurs using the retry module.
DO NOT use this module if you are piping request
instances.
If you are listening to data
events to buffer, beware that you must reset everything when a replay
occurs.
This is why pipping
is not supported.
$ npm install request-replay
var fs = require('fs');
var request = require('request');
var replay = require('request-replay');
// Note that the options argument is optional
// Accepts the same options the retry module does and an additional
// errorCodes array with error codes that cause the replay to happen
// Check out the code to see which is the default value for it
replay(request('http://google.com/doodle.png', { timeout: 10000 }, function (err, response, body) {
// Do things
}), {
retries: 10,
factor: 3
})
.on('socket', function (socket) {
// In some operating systems the socket timeout is 0 so you must explicitly set it
// and close the socket once reached
socket.setTimeout(10000, socket.end.bind(socket));
})
.on('replay', function (replay) {
// "replay" is an object that contains some useful information
console.log('request failed: ' + replay.error.code + ' ' + replay.error.message);
console.log('replay nr: #' + replay.number);
console.log('will retry in: ' + replay.delay + 'ms')
});
Note that the default retry options are modified to be more appropriate for requests:
retries
: The maximum amount of times to retry the operation. Default is5
.factor
: The exponential factor to use. Default is2
.minTimeout
: The amount of time before starting the first retry. Default is2000
.maxTimeout
: The maximum amount of time between two retries. Default is35000
.randomize
: Randomizes the timeouts by multiplying with a factor between1
to2
. Default istrue
.
Released under the MIT License.