A Writable Stream object that uploads to s3 objects, periodically rotating to a new object name.
See also tails3 for a script to tail the log files produced by s3-streamlogger.
npm install --save s3-streamlogger
var S3StreamLogger = require('s3-streamlogger').S3StreamLogger;
var s3stream = new S3StreamLogger({
bucket: "mys3bucket",
access_key_id: "...",
secret_access_key: "..."
});
s3stream.write("hello S3");
npm install --save winston
npm install --save s3-streamlogger
var winston = require('winston');
var S3StreamLogger = require('s3-streamlogger').S3StreamLogger;
var s3_stream = new S3StreamLogger({
bucket: "mys3bucket",
access_key_id: "...",
secret_access_key: "..."
});
var logger = new (winston.Logger)({
transports: [
new (winston.transports.File)({
stream: s3_stream
})
]
});
logger.info('Hello Winston!');
var S3StreamLogger = require('s3-streamlogger').S3StreamLogger;
var s3stream = new S3StreamLogger({
bucket: "mys3bucket",
folder: "my/nested/subfolder",
access_key_id: "...",
secret_access_key: "..."
});
s3stream.write("hello S3");
tails3 expects messages to be logged as json (the default for the file transport), with hostname and (for critical errors), stack properties to each log object, in addition to the standard timestamp, level and message properties. You can provide these using the third "metadata" option to winston's log method:
logger.log(level, message, {hostname: ... , stack: ...});
When there is an error writing to s3, the stream emits an 'error' event with details. You should take care not to log these errors back to the same stream (as that is likely to cause infinite recursion). Instead log them to the console, to a file, or to SNS using winston-sns.
Note that these errors will result in uncaught exceptions unless you have an
error
event handler registered, for example:
s3_stream.on('error', function(err){
// there was an error!
some_other_logging_transport.log('error', 'logging transport error', err)
});
Name of the S3 bucket to upload data to. Must exist.
Can also be provided as the environment variable BUCKET_NAME
.
An optional folder to stream log files to. Takes a path string, eg: "my/subfolder" or "nested".
AWS access key ID, must have putObject permission on the specified bucket. Can
also be provided as the environment variable AWS_SECRET_ACCESS_KEY
, or as any
of the other authentication
methods
supported by the AWS SDK.
AWS secret key for the access_key_id
specified. Can also be provided as the
environment variable AWS_SECRET_KEY_ID
, or as any of the other
authentication
methods
supported by the AWS SDK.
Configuration object for the AWS SDK. The full list of options is available on the AWS SDK Configuration Object page. This is an alternative to using access_key_id and secret_access_key and is overwritten by them if both are used.
Format of file names to create, accepts strftime specifiers. Defaults to "%Y-%m-%d-%H-%M-%S-%L-<current git branch>-<hostname>.log"
. The Date() used to fill the format specifiers is created with the current UTC time, but still has the current timezone, so any specifiers that perform timezone conversion will return incorrect dates.
If you use a format of the form %Y-%m-%d-%H-%M-<stage>-<hostname>.log
, then
you can use tails3 to tail the log files
being generated by S3StreamLogger
.
Files will be rotated every rotate_every
milliseconds. Defaults to 3600000 (60
minutes).
Files will be rotated when they reach max_file_size
bytes. Defaults to 200 KB.
Files will be uploaded every upload_every
milliseconds. Defaults to 20
seconds.
Files will be uploaded if the un-uploaded data exceeds buffer_size
bytes.
Defaults to 10 KB.
The server side encryption AES256
algorithm used when storing objects in S3.
Defaults to false.
ISC: equivalent to 2-clause BSD.