Logfmt Parser Transform

The Vector logfmt_parser transform parses a log field's value in the logfmt format.

Configuration

[transforms.my_transform_id]
# General
type = "logfmt_parser" # required
inputs = ["my-source-or-transform-id"] # required
drop_field = true # optional, default
field = "message" # optional, default
# Types
types.status = "int" # example
types.duration = "float" # example
types.success = "bool" # example
types.timestamp_iso8601 = "timestamp|%F" # example
types.timestamp_custom = "timestamp|%a %b %e %T %Y" # example
types.timestamp_unix = "timestamp|%F %T" # example
types.parent.child = "int" # example
  • commonoptionalbool

    drop_field

    If the specified field should be dropped (removed) after parsing.

    • Default: true
    • View examples
  • commonoptionalstring

    field

    The log field to parse. See Format Specification for more info.

    • Default: "message"
    • View examples
  • commonoptionaltable

    types

    Key/value pairs representing mapped log field names and types. This is used to coerce log fields from strings into their proper types. The available types are listed in the Types list below.

    Timestamp coercions need to be prefaced with timestamp|, for example "timestamp|%F". Timestamp specifiers can use either of the following:

    1. One of the built-in-formats listed in the Timestamp Formats table below.
    2. The time format specifiers from Rust's chrono library.

    Types

    • array
    • bool
    • bytes
    • float
    • int
    • map
    • null
    • timestamp (see the table below for formats)

    Timestamp Formats

    FormatDescriptionExample
    %F %TYYYY-MM-DD HH:MM:SS2020-12-01 02:37:54
    %v %TDD-Mmm-YYYY HH:MM:SS01-Dec-2020 02:37:54
    %FT%TISO 8601[RFC
    3339](https://tools.ietf.org/html/rfc3339) format without time zone

    2020-12-01T02:37:54 %a, %d %b %Y %T | RFC 822/2822 without time zone | Tue, 01 Dec 2020 02:37:54 %a %d %b %T %Y | date command output without time zone | Tue 01 Dec 02:37:54 2020 %a %b %e %T %Y | ctime format | Tue Dec 1 02:37:54 2020 %s | UNIX timestamp | 1606790274 %FT%TZ | ISO 8601/RFC 3339 UTC | 2020-12-01T09:37:54Z %+ | ISO 8601/RFC 3339 UTC with time zone | 2020-12-01T02:37:54-07:00 %a %d %b %T %Z %Y | date command output with time zone | Tue 01 Dec 02:37:54 PST 2020 %a %d %b %T %z %Y| date command output with numeric time zone | Tue 01 Dec 02:37:54 -0700 2020 %a %d %b %T %#z %Y | date command output with numeric time zone (minutes can be missing or present) | Tue 01 Dec 02:37:54 -07 2020

    Note: the examples in this table are for 54 seconds after 2:37 am on December 1st, 2020 in Pacific Standard Time.

Telemetry

This component provides the following metrics that can be retrieved through the internal_metrics source. See the metrics section in the monitoring page for more info.

  • counter

    processing_errors_total

    The total number of processing errors encountered by this component. This metric includes the following tags:

    • component_kind - The Vector component kind.

    • component_name - The Vector component ID.

    • component_type - The Vector component type.

    • error_type - The type of the error

    • instance - The Vector instance identified by host and port.

    • job - The name of the job producing Vector metrics.

  • counter

    processed_events_total

    The total number of events processed by this component. This metric includes the following tags:

    • component_kind - The Vector component kind.

    • component_name - The Vector component ID.

    • component_type - The Vector component type.

    • file - The file that produced the error

    • instance - The Vector instance identified by host and port.

    • job - The name of the job producing Vector metrics.

  • counter

    processed_bytes_total

    The total number of bytes processed by the component. This metric includes the following tags:

    • component_kind - The Vector component kind.

    • component_name - The Vector component ID.

    • component_type - The Vector component type.

    • instance - The Vector instance identified by host and port.

    • job - The name of the job producing Vector metrics.

Examples

Given the following Vector event:

{
"log": {
"message": "at=info method=GET path=/ host=myapp.herokuapp.com request_id=8601b555-6a83-4c12-8269-97c8e32cdb22 fwd=\"204.204.204.204\" dyno=web.1 connect=1ms service=18ms status=200 bytes=13 tls_version=tls1.1 protocol=http"
}
}

And the following configuration:

[transforms.logfmt_parser]
type = "logfmt_parser"
field = "message"
drop_field = true
types.bytes = "int"
types.status = "int"

The following Vector log event will be output:

{
"at": "info",
"method": "GET",
"path": "/",
"host": "myapp.herokuapp.com",
"request_id": "8601b555-6a83-4c12-8269-97c8e32cdb22",
"fwd": "204.204.204.204",
"dyno": "web.1",
"connect": "1ms",
"service": "18ms",
"status": 200,
"bytes": 13,
"tls_version": "tls1.1",
"protocol": "http"
}

How It Works

Format Specification

Logfmt is, unfortunately, a very loosely defined format. There is no official specification for the format and Vector makes a best effort to parse key/value pairs delimited with a =. It works by splitting the field's value on non-quoted white-space and then splitting each token by a non-quoted = character. This makes the parsing process somewhat flexible in that the string does not need to be strictly formatted.

For example, the following log line:

log event
{
"message": "Hello world duration=2s user-agent="Firefox/47.3 Mozilla/5.0""
}

Will be successfully parsed into:

log event
{
"message": "Hello world duration=2s user-agent="Firefox/47.3 Mozilla/5.0"",
"duration": "2s",
"user-agent": "Firefox/47.3 Mozilla/5.0"
}

Key/Value Parsing

This transform can be used for key/value parsing. Logfmt refers to a loosely defined spec that parses a key/value pair delimited by a = character. This section, and it's keywords, is primarily added to assist users in finding this transform for these terms.

Quoting Values

Values can be quoted to capture spaces, and quotes can be escaped with \. For example

key1="value with spaces" key2="value with spaces and \""

Would result in the following log event:

log event
{
"key1": "value with spaces",
"key2": "value with spaces and \""
}