telegraf/docs/PROCESSORS.md

### Processor Plugins

This section is for developers who want to create a new processor plugin.

### Processor Plugin Guidelines

* A processor must conform to the [telegraf.Processor][] interface.
* Processors should call `processors.Add` in their `init` function to register
  themselves.  See below for a quick example.
* To be available within Telegraf itself, plugins must add themselves to the
  `github.com/influxdata/telegraf/plugins/processors/all/all.go` file.
* The `SampleConfig` function should return valid toml that describes how the
  processor can be configured. This is include in the output of `telegraf
  config`.
- The `SampleConfig` function should return valid toml that describes how the
  plugin can be configured. This is included in `telegraf config`.  Please
  consult the [SampleConfig][] page for the latest style guidelines.
* The `Description` function should say in one line what this processor does.
- Follow the recommended [CodeStyle][].

### Processor Plugin Example

```go
package printer

// printer.go

import (
	"fmt"

	"github.com/influxdata/telegraf"
	"github.com/influxdata/telegraf/plugins/processors"
)

type Printer struct {
	Log telegraf.Logger `toml:"-"`
}

var sampleConfig = `
`

func (p *Printer) SampleConfig() string {
	return sampleConfig
}

func (p *Printer) Description() string {
	return "Print all metrics that pass through this filter."
}

// Init is for setup, and validating config.
func (p *Printer) Init() error {
	return nil
}

func (p *Printer) Apply(in ...telegraf.Metric) []telegraf.Metric {
	for _, metric := range in {
		fmt.Println(metric.String())
	}
	return in
}

func init() {
	processors.Add("printer", func() telegraf.Processor {
		return &Printer{}
	})
}
```

### Streaming Processors

Streaming processors are a new processor type available to you. They are
particularly useful to implement processor types that use background processes
or goroutines to process multiple metrics at the same time. Some examples of this
are the execd processor, which pipes metrics out to an external process over stdin
and reads them back over stdout, and the reverse_dns processor, which does reverse
dns lookups on IP addresses in fields. While both of these come with a speed cost,
it would be significantly worse if you had to process one metric completely from
start to finish before handling the next metric, and thus they benefit
significantly from a streaming-pipe approach.

Some differences from classic Processors:

* Streaming processors must conform to the [telegraf.StreamingProcessor][] interface.
* Processors should call `processors.AddStreaming` in their `init` function to register
  themselves.  See below for a quick example.

### Streaming Processor Example

```go
package printer

// printer.go

import (
	"fmt"

	"github.com/influxdata/telegraf"
	"github.com/influxdata/telegraf/plugins/processors"
)

type Printer struct {
	Log telegraf.Logger `toml:"-"`
}

var sampleConfig = `
`

func (p *Printer) SampleConfig() string {
	return sampleConfig
}

func (p *Printer) Description() string {
	return "Print all metrics that pass through this filter."
}

// Init is for setup, and validating config.
func (p *Printer) Init() error {
	return nil
}

// Start is called once when the plugin starts; it is only called once per
// plugin instance, and never in parallel.
// Start should return once it is ready to receive metrics.
// The passed in accumulator is the same as the one passed to Add(), so you
// can choose to save it in the plugin, or use the one received from Add().
func (p *Printer) Start(acc telegraf.Accumulator) error {
}

// Add is called for each metric to be processed. The Add() function does not
// need to wait for the metric to be processed before returning, and it may
// be acceptable to let background goroutine(s) handle the processing if you
// have slow processing you need to do in parallel.
// Keep in mind Add() should not spawn unbounded goroutines, so you may need
// to use a semaphore or pool of workers (eg: reverse_dns plugin does this).
// Metrics you don't want to pass downstream should have metric.Drop() called,
// rather than simply omitting the acc.AddMetric() call
func (p *Printer) Add(metric telegraf.Metric, acc telegraf.Accumulator) error {
	// print!
	fmt.Println(metric.String())
	// pass the metric downstream, or metric.Drop() it.
	// Metric will be dropped if this function returns an error.
	acc.AddMetric(metric)

	return nil
}

// Stop gives you an opportunity to gracefully shut down the processor.
// Once Stop() is called, Add() will not be called any more. If you are using
// goroutines, you should wait for any in-progress metrics to be processed
// before returning from Stop().
// When stop returns, you should no longer be writing metrics to the
// accumulator.
func (p *Printer) Stop() error {
}

func init() {
	processors.AddStreaming("printer", func() telegraf.StreamingProcessor {
		return &Printer{}
	})
}
```

[SampleConfig]: https://github.com/influxdata/telegraf/wiki/SampleConfig
[CodeStyle]: https://github.com/influxdata/telegraf/wiki/CodeStyle
[telegraf.Processor]: https://godoc.org/github.com/influxdata/telegraf#Processor
[telegraf.StreamingProcessor]: https://godoc.org/github.com/influxdata/telegraf#StreamingProcessor