README.md in fluent-plugin-honeycomb-0.5.0 vs README.md in fluent-plugin-honeycomb-0.6.0

@@ -27,9 +27,10 @@
 | `sample_rate` | integer | no | Sample your event stream by sending 1 out of every N events. |
 | `include_tag_key` | bool | no | Whether to include the Fluentd tag in the submitted event. |
 | `tag_key` | string | no | If `include_tag_key` is `true`, the tag key name in the event (default: `fluentd_tag`).
 | `flatten_keys` | array | no | Flatten nested JSON data under these keys into the top-level event.
 | `dataset_from_key` | string | no | Look for this key in each event, and use its value as the destination dataset. If an event doesn't contain the key, it'll be sent to the dataset given by the `dataset` parameter.
+| `presampled_key` | string | no | Look for this key in each event, and use its value as the sample rate for the record. If an event doesn't contain the key, the logic around `sample_rate` will be used (1 out of every N events will be sent instead). If an event's `presampled_key` value is not a positive integer, the value is discarded, and the event is sent to Honeycomb with a `samplerate` of 1 (the default), ignoring the `sample_rate` configuration option. |
 
 ### Buffering options
 
 `fluent-plugin-honeycomb` supports the [standard configuration options](http://docs.fluentd.org/v0.12/articles/buffer-plugin-overview) for buffered output plugins.