Fluent::Plugin::Beacon, a plugin for Fluentd

fluent-plugin-f5-beacon is a buffered output plugin for Fluentd and F5 Beacon.

$ fluent-gem install fluent-plugin-f5-beacon

Alternatively, with td-agent:

$ td-agent-gem install fluent-plugin-f5-beacon

Or via source (swap td-agent-gem for fluent-gem as needed):

$ git clone https://github.com/f5devcentral/fluent-plugin-f5-beacon.git
$ cd fluent-plugin-f5-beacon/
$ fluent-gem build fluent-plugin-f5-beacon.gemspec
$ fluent-gem install fluent-plugin-f5-beacon-#.#.#.gem

Just like other regular output plugins, use type f5-beacon in your Fluentd configuration under match scope:

@type f5_beacon

Options:

source_name: The name of the source within Beacon. This is required.

token: The Beacon ingestion token to be used. This is required.

measurement: The measurement/series to use. The default is nil. If not specified, Fluentd's tag is used.

time_key: Use the value of this tag if it exists in the event instead of the event timestamp.

auto_tags: Enable/disable auto-tagging behavior which makes strings tags. The default is false.

tag_keys: The names of the keys to use as tags instead of fields.

sequence_tag: The name of the tag whose value is incremented for the consecutive simultaneous events and reset to zero for a new event with the different timestamp.

cast_number_to_float: Enable/disable casting numbers to floats. Within a measurement, a given field must be integer or float. If your pipeline cannot unify the record value, this parameter may help avoid errors due to type conflicts.

<match mylog.*>
  @type f5_beacon
  @id f5_beacon
  source_name server1
  token a-123456789#token
  tag_keys ["key1", "key2"]
</match>

The plugin is a buffered output plugin. So additional buffer configuration (with default values) looks like:

<buffer>
  @type memory
  chunk_limit_size 524288 # 512 * 1024
  chunk_limit_records 1024
  flush_interval 60
  retry_limit 17
  retry_wait 1.0
  num_threads 1
</buffer>

Details around buffering can be found here.

bundle exec rake test