An elasticsearch transport for the winston logging toolkit.
- logstash compatible message structure.
- Thus consumable with kibana.
- Date pattern based index names.
- Custom transformer function to transform logged data into a different message structure.
- Querying.
- Real buffering of messages in case of unavailable ES.
npm install --save winston winston-elasticsearch
var winston = require('winston');
var Elasticsearch = require('winston-elasticsearch');
var esTransportOpts = {
level: 'info'
};
winston.add(winston.transports.Elasticsearch, esTransportOpts);
// - or -
var logger = new winston.Logger({
transports: [
new Elasticsearch(esTransportOpts)
]
});
level
[info
] Messages logged with a severity greater or equal to the given one are logged to ES; others are discarded.index
[none] the index to be used. This option is mutually exclusive withindexPrefix
.indexPrefix
[logs
] the prefix to use to generate the index name according to the pattern<indexPrefix>-<indexSuffixPattern>
.indexSuffixPattern
[YYYY.MM.DD
] a Moment.js compatible date/ time pattern.messageType
[log
] the type (path segment after the index path) under which the messages are stored under the index.transformer
[see below] a transformer function to transform logged data into a different message structure.ensureMappingTemplate
[true
] If set totrue
, the givenmappingTemplate
is checked/ uploaded to ES when the module is sending the fist log message to make sure the log messages are mapped in a sensible manner.mappingTemplate
[see fileindex-template-mapping.json
file] the mapping template to be ensured as parsed JSON.flushInterval
[2000
] distance between bulk writes in ms.client
An elasticsearch client instance. If given, all following options are ignored.clientOpts
An object hash passed to the ES client. See its docs for supported options.consistency
[one
] The consistency hint used to store messages in ES. Possible valuesone
,quorum
,all
.
When changing the indexPrefix
and/ or the transformer
,
make sure to provide a matching mappingTemplate
.
The transformer function allows to transform the log data structure as provided by winston into a sturcture more appropriate for indexing in ES.
The default transformer function's transformation is shwon below.
Input:
{
"message": "Some message",
"level": "info",
"meta": {
"method": "GET",
"url": "/sitemap.xml",
...
}
}
}
Output:
{
"@timestamp": "2015-09-30T05:09:08.282Z",
"message": "Some message",
"severity": "info",
"fields": {
"method": "GET",
"url": "/sitemap.xml",
...
}
}
The @timestamp
is generated in the transformer.
Note that in current logstash versions, the only "standard fields" are @timestamp and @version,
anything else ist just free.
A custom trunsformer function can be provided in the options hash.
error
: in case of any error.
An example assuming default settings.
logger.info('Some message', <req meta data>);
The log message generated by this module has the following structure:
{
"@timestamp": "2015-09-30T05:09:08.282Z",
"message": "Some log message",
"severity": "info",
"fields": {
"method": "GET",
"url": "/sitemap.xml",
"headers": {
"host": "www.example.com",
"user-agent": "Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)",
"accept": "*/*",
"accept-encoding": "gzip,deflate",
"from": "googlebot(at)googlebot.com",
"if-modified-since": "Tue, 30 Sep 2015 11:34:56 GMT",
"x-forwarded-for": "66.249.78.19"
}
}
}
This message would be POSTed to the following endpoint:
http://localhost:9200/logs-2015.09.30/log/
So the default mapping uses an index pattern logs-*
.