Cette page n'est pas encore disponible en français, sa traduction est en cours.
Si vous avez des questions ou des retours sur notre projet de traduction actuel, n'hésitez pas à nous contacter.

Observability Pipelines is not available on the US1-FED Datadog site.

Overview

Observability Pipelines Worker configurations can collect, transform, and route your logs from any source to any destination. The configuration file supports YAML, TOML, and JSON. The three main configuration components are sources, transforms, and sinks.

Set up an example source

Source components define how the Observability Pipelines Worker collects or receives data from observability data sources.

Create a YAML configuration file and add the following source example:

sources:
  generate_syslog:
    type: demo_logs
    format: syslog
    count: 100
[sources.generate_syslog]
   type = "demo_logs"
   format = "syslog"
   count = 100
"sources": {
    "generate_syslog": {
      "type": "demo_logs",
      "format": "syslog",
      "count": 100
    }
}

This source component has a unique ID of generate_syslog. This unique ID is important for transforming and routing the data with thesink component.

type is the source type from which the Observability Pipelines Worker collects observability data. This example uses a demo_logs source, which creates sample log data that enables you to simulate different types of events in various formats. The format option tells the demo_logs source which type of logs to emit, in this case, Syslog format. The count option tells the demo_logs source how many lines to emit.

See all supported sources in the Sources documentation.

Set up an example transform

Use the following example to define a transform component that manipulates the data collected from the demo_logs source.

transforms:
  remap_syslog:
    inputs:
      - generate_syslog
    type: remap
    source: |2
        structured = parse_syslog!(.message)
        . = merge(., structured)
[transforms.remap_syslog]
   inputs = ["generate_syslog" ]
   type = "remap"
   source = '''
     structured = parse_syslog!(.message)
     . = merge(., structured)
'''
"transforms": {
    "remap_syslog": {
      "inputs": [
        "generate_syslog"
      ],
      "type": "remap",
      "source": "  structured = parse_syslog!(.message)\n  . = merge(., structured)\n"
    }
  }

In this transforms.remap_syslog component, the inputs option is set to generate_syslog, which means it receives events from the previously defined generate_syslog source. The transform’s component type is remap.

The source contains the list of remapping transformations to apply to each event that the Observability Pipelines Worker receives. In this example, only one operation, parse_syslog, is performed, but multiple operations can be added.

The parse_syslog function receives a single field called message, which contains the Syslog event that is generated in the generate_syslog source. This function parses the content of the Syslog-formatted message and emits it as a structured event.

This transform example showcases only a portion of the Observability Pipelines Worker’s ability to shape and transform your data*. See the Transforms documentation for all supported transforms, ranging from sampling, filtering, enrichment, and more.

Set up an example sink

With the data parsed in the transform component, use the following sink example to route the data to a destination.

sinks:
  emit_syslog:
    inputs:
      - remap_syslog
    type: console
    encoding:
      codec: json
[sinks.emit_syslog]
inputs = [ "remap_syslog" ]
type = "console"

  [sinks.emit_syslog.encoding]
  codec = "json"
"sinks": {
    "emit_syslog": {
      "inputs": [
        "remap_syslog"
      ],
      "type": "console",
      "encoding": {
        "codec": "json"
      }
    }
}

This sink (or destination) component has the ID of emit_syslog. The inputs option specifies that the events generated by the remap_syslog transform are processed with this sink. The encoding option tells the sink to emit the events in JSON format.

See the Sinks documentation for all supported sinks.

Put it all together

With these three basic components, a source, transform, and sink, you now have a working Observability Pipelines configuration file.

sources:
  generate_syslog:
    type: demo_logs
    format: syslog
    count: 100
transforms:
  remap_syslog:
    inputs:
      - generate_syslog
    type: remap
    source: |2
        structured = parse_syslog!(.message)
        . = merge(., structured)

sinks:
  emit_syslog:
    inputs:
      - remap_syslog
    type: console
    encoding:
      codec: json
[sources.generate_syslog]
type = "demo_logs"
format = "syslog"
count = 100

[transforms.remap_syslog]
inputs = [ "generate_syslog" ]
type = "remap"
source = '''
  structured = parse_syslog!(.message)
  . = merge(., structured)
'''

[sinks.emit_syslog]
inputs = [ "remap_syslog" ]
type = "console"

  [sinks.emit_syslog.encoding]
  codec = "json"
{
  "sources": {
    "generate_syslog": {
      "type": "demo_logs",
      "format": "syslog",
      "count": 100
    }
  },
  "transforms": {
    "remap_syslog": {
      "inputs": [
        "generate_syslog"
      ],
      "type": "remap",
      "source": "  structured = parse_syslog!(.message)\n  . = merge(., structured)\n"
    }
  },
  "sinks": {
    "emit_syslog": {
      "inputs": [
        "remap_syslog"
      ],
      "type": "console",
      "encoding": {
        "codec": "json"
      }
    }
  }
}

Run the following command to compile and run this configuration:

vector --config ./<configuration_filename>

If successfully setup, the parsed demo logs are printed in JSON format.

Further Reading