A pino transport for filtering log entries and writing them to a file that automatically rotates. Log entries that are not filtered continue out of stdout.
This is useful for logging when you want different entries in different rotating files. Using rotating files allows you to roll logs across multiple files so that individual file sizes don't grow too large and old files get deleted.
At its core, this processor takes a simple config js file that contains an output path and additional options for the rotation. Under the hood, it uses the rotating-file-stream - check out the options. It also has a filter function to determine if the log entry should go to the log file, or be forwarded through to stdout.
Install this module as a dependency in your project with npm install @vrbo/pino-rotating-file
After this, you can reference it as rotate-logs in your npm scripts in your package.json file when running your main application, like this:
...
"scripts": {
"start": "node ./index.js | rotate-logs --config=.request-log.js"
}
...module.exports = {
filter(data) {return !!data.req},
output: {
path: "request.log", // name of file
options: {
path: "logs/", // path to write files to
size: "10M", // max file size
rotate: 5 // keep 5 rotated logs
}
}
}If no
--configflag is given, it will look for the file.rotate.jsby default
You can chain all these together to split your logs into separate, automatically rotating log files:
...
"scripts": {
"start": "node ./index.js | rotate-logs --config=.request-log.js | rotate-logs --config=.error-log.js | rotate-logs --config=.server-log.js"
}
...Where .request-log.js pulls all the logs that have req tags, .error-log.js pulls all the logs with error tags, and .server-log.js puts all the rest into another file.
NOTE: if the config provided does not exist or fails to load, then all logs will be passed to stdout.