Skip to content

Latest commit

 

History

History
105 lines (72 loc) · 5.19 KB

README.md

File metadata and controls

105 lines (72 loc) · 5.19 KB
Raw

Splunk REST endpoint examples

Overview

This app contains several REST endpoints that can be used as a starting point for Splunk app developers.

Installation

To install, clone the repository into the $SPLUNK_HOME/etc/apps directory and restart Splunk.

Overview

A Splunk REST handler is a script that is invoked when a request is made to Splunk's REST API at a specific URI. By setting configurations in the restmap.conf file of your app, and writing the corresponding code, you can enable Splunk to execute code of your choice in response to an HTTP request.

REST handlers are a powerful method for executing processes under the control of the primary splunkd process. They complement the other mechanisms provided by Splunk for external process execution, which include: modular inputs, CherryPy controllers, and custom search commands.

Interfaces

Splunk provides several interfaces that can be extended for custom REST API development.

  • "Old-style" handlers derived from the rest.BaseRestHandler class
  • EAI handlers derived from the admin.MConfigHandler class
  • "New-style" handlers derived from the PersistentServerConnectionApplication class. These are often referred to as "persistent" handlers; however this is slightly misleading, as noted below.

This app does NOT provide an example of a class derived from rest.BaseRestHandler; in nearly all cases, it is preferable to derive from PersistentServerConnectionApplication for non-EAI use cases.

EAI (Extensible Administration Interface)

The Extensible Administration Interface is the interface used by Splunk Core developers to easily extend the user interface to accommodate new adminstrative capabilities. The saved searches administration UI is a (complex) example of an interface driven by an EAI handler. The EAI interface provides certain services such as pagination, role-based access control, and namespacing to the developer; however, for the developer, writing an EAI handler is generally more complex than writing a "new-style" handler. EAI handlers can provide output in Atom feed format or (in more recent Splunk versions) as JSON.

Persistence

Prior to the introduction of "persistent mode" handlers, an external process - usually a Python interpreter - would be instantiated for every REST API call.

In Splunk 6.4, a new persistent mode has been added to REST handler interfaces enabling them to serve multiple REST API calls in a single process execution. This vastly reduces the number of processes that must be created to handle API traffic, and brings REST handlers roughly in line with the performance of CherryPy controllers.

Note that a handler in persistent mode will exit after it has been completely idle for a period of 60 seconds. This timeout is not currently configurable.

While the "new-style" protocol is commonly referred to as the "persistent" protocol, EAI handlers can also be run in persistent mode. An important distinction is that while a persistent handler written using the new protocol cannot be trivially converted to a non-persistent handler, EAI handlers can be "toggled" between persistent and non-persistent mode transparently, with no code changes required.

App Contents

  1. Simple persistent handler: echo_persistent_handler.py

    This handler is the simplest possible demonstration of the new "raw" REST interface. It echoes all arguments passed into it as output. Usage examples may be found at echo_persistent_examples.

  2. Simple EAI handler: simple_eai_handler.py

    This handler is the simplest possible EAI handler, provided solely for the purpose of demonstrating the mechanism for populating EAI output. It is unlikely to be useful for any serious purpose. Usage examples may be found at simple_eai_examples.

  3. Persistent handler demonstrating input parameters and output formatting options: multiformat_persistent_handler.py.

Usage via splunkd

Access to custom REST handlers via splunkd, which listens on port 8089, can be performed from localhost using the "curl" utility, similar to any other Splunk REST endpoint.

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

Usage via splunkweb

A REST endpoint can be made accessible on the Splunkweb port (8000) via use of an "expose" stanza in web.conf.

Requests to port 8000 must circumvent CSRF protection restriction. For command-line testing, the simplest way to obtain a token is to log in to Splunk using Chrome and, using Chrome's developer tools, right-click on any request and select "Copy as cURL".

A new REST request to port 8000 can then be crafted by replacing the parameters shown in <brackets>:

curl -k https://localhost:8000/en-US/splunkd/__raw/services/echo_persistent \
     -H 'Cookie: session_id_8000=<session_id>; splunkd_8000=<splunkd_8000>; splunkweb_csrf_token_8000=<csrf_token>'

Note that the URL should be rooted at /splunkd/__raw as shown above.

License

No license, warranty, or claim of correctness for this code is provided, promised, or implied. This code is provided solely for educational purposes and may not reflect the actual state of Splunk's product line. Please direct all complaints to the author.