Creating searches using the REST API

Use the search/jobs endpoint to create a search job in a Splunk deployment. However, before creating searches you should be aware of how searches work and how to structure a search so you can easily access the results.

Learn about searches

You can learn about searches and how to write them in the Search Manual.

Here are some highlights from the Search Reference that can help you get started creating searches:

• Splunk Enterprise Quick Reference Guide Provides fundamental search concepts, commands, functions, and examples.

• Search command by category Find the commands you need in this list, organized by category.

• Command reference Complete syntax of Splunk search commands.

Additional information about working with searches is in the Search Manual:

• Write better searches Provides some simple rules of thumb and best practices to help you write searches that run more efficiently.
• Execution costs Describes how to troubleshoot the efficiency of a search.
• View search job properties Lists and defines the properties of a search job.

REST endpoints for searches

Here is a brief description of some of the key endpoints for creating and accessing searches.

/search/jobs
Create searches or access the results of search jobs. Returns a search ID (sid) that you use when accessing the results of a search.

/search/jobs/export
Stream search results as they become available. Does not create a search ID for later access.

/search/jobs/{search_id}/control
Execute a job control command for a search, such as pause, setpriority, or finalize.

/search/jobs/{search_id}/events
Return untransformed events of a search.

/search/jobs/{search_id}/results
Return transformed events of a search.

/search/jobs/{search_id}/summary
Return summary information for fields of a search.

/search/jobs/{search_id}/timeline
Return event distribution over time of the so-far-read untransformed events.

/saved/searches
Create or access the configuration of saved searches.

Tips on creating searches

When creating a search (POST /search/jobs), consider the following properties of the search:

max_count
Set this parameter for searches returning more than the default maximum of 10000 events. Otherwise you may not be able to retrieve results in excess of the default.

status_buckets
To access summary and timeline information from a search job, specify a value for status_buckets. The default value is zero. For example, searches spawned from the Splunk timeline specify status_buckets=300.

rf
Use the rf parameter to add required fields to a search. Adding fields guarantees results for the endpoints that return events and a summary. (The required_fields parameter has been deprecated in favor of the rf parameter.)

Tips on accessing searches

When accessing the results of a search (GET /search/jobs/{search_id}), consider the following:

search, offset, and count parameters
Use these parameters to a GET operation to filter or limit the results returned.

dispatchState
dispatchState is one of the properties returned when accessing a search. It provides the state of a search, which can be any of the following:

QUEUED

PARSING

RUNNING

PAUSED

FINALIZING

FAILED

DONE

search job properties
The GET operation for /search/jobs returns all the properties of a search. These properties are described in the parameters to the POST operation. Search job properties are also described in View search job properties in the Search Manual.

performance
The GET operation for /search/jobs returns information that helps you troubleshoot the efficiency of a search. See the "Execution costs" section in View search job properties in the Search Manual.

Example: Create a search

Many calls to Splunk's API involve running some kind of search. For example, you may wish to run a search within Splunk Enterprise and POST the results to a third party application. Use the endpoints located at the ../services/search/<endpoint> URIs.

When you run a search, the search process launches asynchronously. You can poll the jobs or events endpoint to see if your search has finished.

Create a search job

Create a search job using the POST operation at search/jobs/. Set your search as the POST payload. For example:

curl -u admin:changeme -k https://localhost:8089/services/search/jobs -d search="search *"

This simple example runs the search for *. It returns an XML response such as:

<?xml version='1.0' encoding='UTF-8'?>
<response>
  <sid>1258421375.19</sid>
</response>

You need the search ID to retrieve the search, which is returned within the <sid> tags. In the example above this is 1258421375.19.

Check status of a search

Check the status of a search job by accessing the GET operation of search/jobs/. If you know the search's ID, you can access search/jobs/{search_id} to get information about that search only:

curl -u admin:changeme -k https://localhost:8089/services/search/jobs/1258421375.19 

If you're not sure what searches you're running, the GET operation at search/jobs endpoint returns a list of searches with their search IDs.

curl -u admin:changeme -k https://localhost:8089/services/search/jobs/ 

Get search results

Use the results endpoint located at /search/jobs/<sid>/results/ to retrieve your search results. This endpoint returns results only when your search has completed. You can also get output from the events endpoint located at /search/jobs/{search_id}/events/ while your search is still running. For complete search results, use the results endpoint.

You can return search results in JSON, CSV or XML by setting the output_mode parameter. By default, results are returned in XML format.

For example, to retrieve search results in CSV format, make the following call.

Note: The curl listing includes --get because you are passing a parameter to a GET operation.

curl -u admin:changeme \
     -k https://localhost:8089/services/search/jobs/1258421375.19/results/ \
     --get -d output_mode=csv

Note: This is one method that you can use to export large numbers of search results. For more information about exporting search results, as well as information about the other export methods, see "Export search results" in the Search Manual.

Python example

This example script authenticates against a Splunk server and runs a search query in Python. After running the search, the script returns the search ID (sid).

#!/usr/bin/python -u

import urllib
import httplib2
from xml.dom import minidom

baseurl = 'https://localhost:8089'
userName = 'admin'
password = 'password'

searchQuery = '| inputcsv foo.csv | where sourcetype=access_common | head 5'

# Authenticate with server.
# Disable SSL cert validation. Splunk certs are self-signed.
serverContent = httplib2.Http(disable_ssl_certificate_validation=True).request(baseurl + '/services/auth/login',
    'POST', headers={}, body=urllib.urlencode({'username':userName, 'password':password}))[1]

sessionKey = minidom.parseString(serverContent).getElementsByTagName('sessionKey')[0].childNodes[0].nodeValue

# Remove leading and trailing whitespace from the search
searchQuery = searchQuery.strip()

# If the query doesn't already start with the 'search' operator or another 
# generating command (e.g. "| inputcsv"), then prepend "search " to it.
if not (searchQuery.startswith('search') or searchQuery.startswith("|")):
    searchQuery = 'search ' + searchQuery

print searchQuery

# Run the search.
# Again, disable SSL cert validation. 
print httplib2.Http(disable_ssl_certificate_validation=True).request(baseurl + '/services/search/jobs','POST',
    headers={'Authorization': 'Splunk %s' % sessionKey},body=urllib.urlencode({'search': searchQuery}))[1]