Golang Hook for the Logrus logging library, in order to output logs to InfluxDB.
go get github.com/stevegood/influxus
import (
"github.com/sirupsen/logrus"
"github.com/influxdata/influxdb/client/v2"
"github.com/stevegood/influxus"
)
func init() {
// Create the InfluxDB client.
influxClient, err := client.NewHTTPClient(client.HTTPConfig{
Addr: "http://localhost:8086",
})
if err != nil {
logrus.Fatalf("Error while creating the client: %v", err)
}
// Create and add the hook.
hook, err := influxus.NewHook(
&influxus.Config{
Client: influxClient,
Database: "logrus", // DATABASE MUST BE CREATED
DefaultMeasurement: "logrus",
BatchSize: 1, // default is 100
BatchInterval: 1, // default is 5 seconds
})
if err != nil {
logrus.Fatalf("Error while creating the hook: %v", err)
}
// Add the hook to the standard logger.
logrus.StandardLogger().Hooks.Add(hook)
}We use a non-blocking model for the hook. Each time an entry is fired, we create an InfluxDB point and put that through a buffered channel. It is then picked up by a worker goroutine that will handle the current batch and write everything once the batch hit the size or time limit.
We chose to take in as a parameter the InfluxDB client so that we have greater flexibility over the way we connect to the Influx Server. All the defaults and configuration options can be found in the GoDoc.
The database should have been previously created. This is by design, as we want to avoid people generating lots of databases. When passing an empty string for the InfluxDB database name, we default to "logrus" as the database name.
We will insert your message into InfluxDB with the field message.
- Concurrent, non-blocking design.
- Add unit tests
- Set up continous integration.