Automate indexed field extractions with HTTP Event Collector

When indexes data, it parses the data stream into a series of events. As part of this process, it adds a number of fields to the event data. These fields include default fields that it adds automatically and any custom fields that you specify. The process of adding fields to events is known as field extraction. There are two types of field extraction: search-time field extraction and indexed field extraction. Indexed fields are incorporated into the index at index time and become part of the event data.

Indexed field extraction doesn't work with data sent to the services/collector/raw endpoint. For more information, see services/collector/raw in the Splunk Enterprise REST API Reference Manual.

Form HEC requests to trigger indexed field extractions

You can trigger indexed extractions of JavaScript Object Notation (JSON) fields in two ways: as part of the main event data or as separate from the event data but still associated with the event.

Use nested JSON inside the event property

Assign the event property, which is at the top level of the JSON being sent to HEC, to a JSON object that contains the custom fields you want to index as key-value pairs. For example, the following event property, from within an HTTP request sent to HEC, specifies two custom fields: club and wins.

"event": {"club":"glee", "wins":["regionals","nationals"]}

In this example, the wins property is a multi-value JSON array. The wins field is assigned both the values in the array.

At the same level as the event property, you must also include a sourcetype property, and set it to a source type that has indexed extraction enabled. You can use any source type that has the INDEXED_EXTRACTIONS setting configured to JSON in the props.conf configuration file, including built-in source types such as _json. See the following example:

"sourcetype":"_json"

Following is an example cURL command that sends an event to HEC on a instance. In this case, the event data contains two custom fields that are extracted at index time:

# Extracting JSON fields
curl https://mysplunkserver.example.com:8088/services/collector -H "Authorization: Splunk 12345678-1234-1234-1234-1234567890AB" -d '{"sourcetype": "_json", "event": {"club":"glee", "wins":["regionals","nationals"]}}'

Add a fields property at the top JSON level

Include the fields property at the top level of the JSON that you send to HEC, which is at the same level as the event property. Adding this property specifies explicit custom fields that are separate from the main event data. This method is useful if you don't want to include the custom fields with the event data, but you want to annotate the data with some extra information, such as where it came from. Using this method is also typically faster than the nested JSON method.

Be aware that you must send HEC requests containing the fields property to the /collector/event endpoint. Otherwise, they aren't indexed.

Assign the fields property to a JSON object that contains the custom fields to be indexed as key-value pairs. For example, the following fields property, from within an HTTP request sent to the Splunk platform instance, specifies two custom fields:club and wins.

"fields": {"club":"glee", "wins":["regionals","nationals"]}

In this example, the wins property is set to a multi-value JSON array. The wins field is assigned both values in the array.

At the same level as the event and fields properties, you must also include a sourcetype property and set it to a source type that has indexed extractions enabled. You can use any source type that has the INDEXED_EXTRACTIONS setting configured to JSON in the props.conf file, including built-in source types such as _json. See the following example:

"sourcetype":"_json"

Following is an example cURL command that sends an event to HEC on a instance. In this case, the event data contains two custom fields that will be extracted at index time:

curl http://mysplunkserver.example.com:8088/services/collector/event -H "Authorization: Splunk 12345678-1234-1234-1234-1234567890AB" -d '{"event": {"my_event": "Hello, McKinley High!"}, "sourcetype": "_json", "fields": {"club":"glee", "wins":["regionals","nationals"]}}'

Only strings can be used as field values.

Search for index-extracted fields

After the data is indexed, you can search for this event using a double-colon ( :: ) indexed extraction notation, as shown here:

sourcetype=_json club::glee

For more information about using extracted fields to retrieve events, see Use fields to retrieve events in the Splunk Enterprise Search Manual.