search

puppetlabs / puppetlabs-splunk_hec

Apache License 2.0
9 stars 25 forks source link
module supported

readme

splunk_hec

Table of Contents
  1. Overview
  2. Requirements
  3. Installation
  4. Custom Installation
  5. SSL Configuration
  6. Fact Configuration
  7. Customized Reporting
  8. Tasks
  9. Advanced Settings
  10. FIPS Mode
  11. Advanced Topics
  12. Known Issues
  13. Breaking Changes
  14. Release Process

Overview

This module provides three services to Puppet and Splunk users.

  1. A report processor to allow sending Puppet Agent run reports to Splunk. When a Puppet agent completes a run and submits some of the report data to Puppet, this module's processor can be invoked to send that report to Splunk. After this module is installed in your environment, to enable sending node reports to Splunk, do the following:

    • Classify your Puppet Servers with the splunk_hec class.
    • Set the url parameter which refers to your Splunk Hec url.
    • Set the token parameter which will be the HEC token you create in Splunk.
    • Set the enable_reports parameter to true.

    For more advanced configuration options including sending reports based on specific conditions see the Customized Reporting section below.

  2. A fact terminus to submit node facts to Splunk. See Fact Terminus Support for details.

  3. A PE Event Forwarding processor for sending data received from the PE Event Forwarding module to Splunk. This data can include the details of Task and Plan executions that were initiated by the Puppet Orchestrator (clicking execute task|plan from the console or puppet command line), or it can be events from rbac, the node classifier, the console, or code-manager. To enable this feature, after the PE Event Forwarding module has been configured, set the events_reporting_enabled parameter on the splunk_hec class to true.

    Note: This is a PE only feature and depends on the PE Event Forwarding module being installed and classified on the Puppet Server nodes in your environment. Please see the documentation in that module for details on how to install and configure that module.

There is also the Puppet Alert Actions app, which contains the alert actions that were previously shipped in the Puppet Report Viewer:

The Puppet Alert Actions app allows you to run custom Tasks in Puppet Enterprise or retrieve detailed Report information about a particular Puppet Event that would be sent to the Puppet Report Viewer. For additional information on configuring Puppet Alert Actions, please see our documentation located here.

There are two Tasks included in this module, splunk_hec:bolt_apply and splunk_hec:bolt_result, that provide a pre-packaged way to compose Bolt Plans that submit data to Splunk every time they are run. Example plans are included which demonstrate task usage.

Requirements

This was tested on the Puppet Enterprise LTS release, Puppet 7 and Puppet 8, using stock gems of yaml, json, and net::https.

Installation

Instructions assume you are using Puppet Enterprise. For Open Source Puppet installations please see the Custom Installation section.

  1. Install the Puppet Report Viewer app in Splunk if not already installed.

  2. Create an HEC token in Splunk:

    • Navigate to Settings > Data Input in your Splunk console.
    • Add a new HTTP Event Collector with a name of your choice.
    • Ensure indexer acknowledgement is not enabled.
    • Click Next and set the source type to Automatic.
    • Ensure the App Context is set to Puppet Report Viewer.
    • Add the main index
    • Set the Default Index to main.
    • Click Review and then Submit.
    • When complete the HEC token should look something like this:

    hec_token

  3. Install the module and add the splunk_hec class to the PE Master node group.

    • Install the splunk_hec module on your Puppet Primary Server.

      • puppet module install puppetlabs-splunk_hec

    • In the PE console, navigate to Node groups and expand PE Infrastructure.

    • Select PE Master and navigate to the Classes tab.

    • Add the splunk_hec class.

    • Configure the following parameters:

      enable_reports           = true
manage_routes            = true
token                    = <TOKEN GENERATED IN STEP 2>
url                      = something like https://splunk-8.splunk.internal:8088/services/collector

    • Commit the changes.

    • Run Puppet on the node group; this will cause a restart of the pe-puppetserver service.

  4. Log into the Splunk console and search index=* sourcetype=puppet:summary, if everything was done properly you should see the reports from the systems in your Puppet environment.

Source Types

  1. puppet:summary

    Puppet agent node reports.

  2. puppet:facts

    Node facts sent by the facts terminus enabled by setting manage_routes to true.

  3. puppet:jobs

    Events gathered from the Puppet Jobs API

The following source types all refer to different types of events gathered from the Puppet Activities API

  1. puppet:activities_rbac

    RBAC events such as creating and/or modifying users or groups, and user logins.

    Note: RBAC events can be disabled from the pe_event_forwarding module for performance reasons. Ensure the disable_rbac parameter is set to false in the pe_event_forwarding module if you wish to send RBAC events to Splunk.

  2. puppet:activities_classifier

    Classifier events such as creating node groups, or modifying the properties of node groups.

  3. puppet:activities_console

    Console events such as requesting Task or Plan runs via the console.

  4. puppet:activities_code_manager

    Code manager events.

Custom Installation

Please Note: If you are installing this module using a control-repo you must have splunk_hec in your production environment's Puppetfile so the Puppet Server process can properly load the required libraries. You can then create a feature branch to enable them and test the configuration, but the libraries must be in production; otherwise the feature branch won't work as expected. If your Puppet Server is in a different environment, please add this module to the Puppetfile in that environment as well.

The steps below will help install and troubleshoot the report processor on a standard Puppet Primary Server; including manual steps to configure compilers (Puppet Servers), and to use the included splunk_hec class. Because one is modifying production machines, these steps allow you to validate your settings before deploying the changes live.

  1. Install the Puppet Report Viewer app in Splunk. This will import the needed source types to configure Splunk's HTTP Endpoint Collector (HEC) and provide a dashboard that will show the reports once they are sent to Splunk.

  2. Create a Splunk HEC Token or use an existing one that sends to main index and does not have acknowledgement enabled. Follow the steps provided by Splunk's Getting Data In Guide if you are new to HTTP Endpoint Collectors.

  3. Install this Puppet module in the environment that your Puppet Servers are using (e.g. production).

  4. Run puppet plugin download on your Puppet Servers to sync the content. Some users with strict permissions may need to run umask 022 first.

    • Please Note: If permissions are too restrictive you may see the following error in the Puppet Server logs:

      Could not find terminus splunk_hec for indirection facts

  5. Create /etc/puppetlabs/puppet/splunk_hec.yaml (see the examples directory), adding your Splunk Server URL to the url parameter (e.g. https://splunk-dev:8088/services/collector) and HEC Token created during step 2 to the splunk_token parameter.

    • You can add timeout as an optional parameter. The default value is 1 second for both open and read sessions, so take value x2 for real world use.

    • PE Only: Provide the pe_console parameter value. This is the FQDN for the PE console, which Splunk can use to lookup further information if the installation utilizes compilers (it is best practice to set this if you're anticipating scaling the installation in the future).

      ---
"url" : "https://splunk-dev.testing.local:8088/services/collector"
"token" : "13311780-EC29-4DD0-A796-9F0CDC56F2AD"

      (Note: If Disaster Recovery is enabled you will need to ensure these settings exist on the Replica node as well. This is often done through the PE HA Replica node group.)

  6. Run puppet apply -e 'notify { "hello world": }' --reports=splunk_hec from the Puppet Server, this will load the report processor and test your configuration settings without actually modifying your Puppet Server's running configuration. If you are using the Puppet Report Viewer app in Splunk then you will see the page update with new data. If not, perform a search by the sourcetype you provided with your HEC configuration.

  7. If configured properly the Puppet Report Viewer app in Splunk will show 1 node in the Overview tab.

  8. Now it is time to roll these settings out to the fleet of Puppet Servers in the installation. For PE users:

    • In the PE console, navigate to Node groups and expand PE Infrastructure.
    • Select PE Master and navigate to the Classes tab.
    • Click Refresh to ensure that the splunk_hec class is loaded.
    • Add new class splunk_hec.
    • From the Parameter drop down list you will need to configure at least url and token, providing the same values from the testing configuration file.
      • Optionally set enable_reports to true if there isn't another component managing the servers reports setting. Otherwise manually add splunk_hec to the settings as described in the manual steps below.
    • Commit changes and run Puppet. It is best to navigate to the PE Certificate Authority node group and run Puppet there first, before running Puppet on the remaining nodes.

  9. For Inventory support in the Puppet Report Viewer, see Fact Terminus Support.

Manual Steps:

SSL Configuration

Configuring SSL support for this report processor and tasks requires that the Splunk HEC service being used has a properly configured SSL certificate. Once the HEC service has a valid SSL certificate, the CA will need to be made available to the report processor to load. The supported path is to install a copy of the Splunk CA to a directory called /etc/puppetlabs/puppet/splunk_hec/ and provide the file name to splunk_hec class.

You can manually update the splunk_hec.yaml file with these settings:

"ssl_ca" : "splunk_ca.cert"

Alternatively, you can create a profile class that copies the splunk_ca.cert as part of invoking the splunk_hec class:

class profile::splunk_hec {
  file { '/etc/puppetlabs/puppet/splunk_hec/splunk_ca.cert':
    ensure => file,
    owner  => 'pe-puppet',
    group  => 'pe-puppet',
    mode   => '0644',
    source => 'puppet:///modules/profile/splunk_hec/splunk_ca.cert',
  }
}

Note: When splunk_hec::ssl_ca is configured it takes precendence over the system certificate store. The report processor will build its own cert store with the provided CA certificate to validate requests against the Splunk HEC endpoint. Alternatively, you can add the Splunk HEC CA to the system certificate store and set splunk_hec::include_system_cert_store to true. This will allow the code to ONLY use the system default store to perform certificate validation.

Fact Configuration

The following parameters are utilized to configure which facts (including custom facts) you would like to send to Splunk:

To configure which facts to collect add the facts_allowlist parameter to the splunk_hec class and modify the array of facts presented.

PE Event Forwarding

PE Customers can install the puppetlabs-pe_event_forwarding module to gather events from the Puppet Orchestrator API and from the Activities API, and then use this module to process that data and send it to Splunk. To enable this feature in a standard installation where this module is already classified to a Puppet Server node and sending reports to Splunk:

  1. See the documenation for puppetlabs-pe_event_forwarding for details on installing and configuring that module. That module will need to be installed and configured before moving on to the next step.
  2. Set the events_reporting_enabled parameter to true.

By default the event_types parameter is configured to send all event types. You can choose which event types to send by setting this parameter to one or more of orchestrator, rbac, classifier, pe-console, or code-manager.

Filtering Event Data

To filter the event data, one can set the following parameters:

The default (no filter set) will send all the data received from the event type. The filters must begin with the top level keys of the event data. One can look at the data in Splunk to see/determine what the top level keys are in the event data.

The format of setting these filters is an array of strings and within the string, you separate the different properties of a single path with a dot . and continue till the desired value. Here's an example of a correctly constructed filter: ['options.scope.nodes', 'report.id', 'environment.name']

NOTE:

Sending from Non Server Nodes

This feature can be configured to send these events from non server nodes if needed. To do this, on the chosen server:

  1. Classify and configure the puppetlabs-pe_event_forwarding according to that module's documentation.

  2. Classify this module with the following parameter values:

    class {'splunk_hec':
  events_reporting_enabled => true,
  url                      => "http://<splunk server name>:8088/services/collector/event",
  token                    => '<splunk token>'
}

    Note: This manifest shows an end point with no SSL protection. To do SSL validation with this module you will have to do all of the steps detailed in the SSL Configuration section, but ensuring you copy the certificate to the correct location on the chosen server where you are classifying splunk_hec and pe_event_forwarding, not the Puppet Server.

Supported Puppet Versions For Event Forwarding

The puppetlabs-pe_event_forwarding module that this feature depends on requires PE versions 2021.2 and above to access the API methods for the module to function properly.

Customized Reporting

As of 0.8.0 and later the report processor can be configured to include Logs and Resource Events along with the existing summary data. Because this data varies between runs and agents in Puppet, it is difficult to predict how much data you will use in Splunk as a result. However, this removes the need for configuring the Detailed Report Generation alerts in Splunk to retrieve that information; which may be useful for large installations that need to retrieve a large amount of data. You can now just send the information from Puppet directly.

Add one or more of these parameters based on the desired outcome, these apply to the state of the puppet runs. You cannot filter by facts on which nodes these are in effect for. As such, you can get logs when a puppet run fails, but not logs when a windows server puppet run fails.

By default this type of reporting is not enabled.

Parameters:

event_types (Requires puppetlabs-pe_event_forwarding module)

Array: Determines which event types should be forwarded to Splunk. Default value includes all event types. This can be one, or any of the following:

include_logs_status

Array: Determines if logs should be included based on the return status of the puppet agent run. This can be none, one, or any of the following:

include_logs_catalog_failure

Boolean: Include logs if a catalog fails to compile. This is a more specific type of failure that indicates a server-side issue.

include_logs_corrective_change

Boolean: Include logs if a there is a corrective change (a PE only feature) - indicating drift was detected from the last time puppet ran on the system.

include_resources_status

Array: Determines if resource events should be included based on the return status of the puppet agent run. Note: This only includes resources whose status is not unchanged - not the entire catalog. The can be none, one, or any of the following:

include_resources_corrective_change

Boolean: Include resource events if a there is a corrective change (a PE only feature) - indicating drift was detected from the last time puppet ran on the system.

only_changes

Boolean: Only process reports when the report status indicates there were changes.

summary_resources_format

String: If include_resources_corrective_change or include_resources_status is set and therefore resource_events are being sent as part of puppet:summary events, we can choose what format they should be sent in. Depending on your usage within Splunk, different formats may be preferable. The possible values are:

Here is an example of the data that will be forwarded to Splunk in each instance:

hash:

{
  "resource_events": {
    "File[/etc/something.conf]": {
      "resource": "File[/etc/something.conf]",
      "failed": false,
      "out_of_sync": true
    }
  }
}

array:

{
  "resource_events": [
    {
      "resource": "File[/etc/something.conf]",
      "failed": false,
      "out_of_sync": true
    }
  ]
}

Tasks

Two tasks are provided for submitting data from a Bolt plan to Splunk. For clarity, we recommend using a different HEC token to distinguish between events from Puppet runs and those generated by Bolt. The Puppet Report Viewer app includes a puppet:bolt sourcetype to faciltate this. Currently SSL validation for Bolt communications to Splunk is not supported.

To setup, add the splunk_hec endpoint as a remote target in your inventory.yaml file:

---
nodes:
  - name: splunk_bolt_hec
    config:
      transport: remote
      remote:
        hostname: <hostname>
        token: <token>
        port: 8088

See the plans directory for working examples of apply and result usage.

Advanced Settings

The splunk_hec module also supports customizing the fact_terminus and facts_cache_terminus names in the custom splunk_routes.yaml it deploys. If you are using a different facts_terminus (i.e. not PuppetDB), you will want to set that parameter.

If you are already using a custom splunk_routes.yaml, these are the equivalent instructions of what the splunk_hec module does. The most important setting is configuring cache: splunk_hec.

FIPS Mode

This module has some limitations in a FIPS environment. In particular, the SSL configuration steps and the available parameters are different under FIPS.

The CA certificate PEM file must be appended to the end of the localcacert file. Find the path to the file by running puppet config print localcacert. Keep in mind that this file will be overwritten any time the puppetserver is upgraded to a new version and this step will have to be done again. Consider copying a backup of this file to a safe location before attempting to add content to it until correct functioning of the Puppet Server and this module can be validated.

  ca_file=$(puppet config print localcacert)
  cp $ca_file ~/$ca_file
  cat my_splunk_hec_ca.pem >> $ca_file

The module must use a different FIPS compliant HTTP client. This client currently lacks support for a number of configurable options. For example, none of the timeout parameters will have any effect. Additionally, ignoring the system certificate store is the default behavior, so there is no need to use the ignore_system_certificate_store parameter. When running in a FIPS environment the following optional parameters are available:

Note: These parameters only have an effect on metrics and will be ignored by the report processor when sending reports and facts.

fips_crl_check

Boolean: In FIPS mode, the HTTP Client will attempt to check the Splunk CA against the Splunk CRL. Unless the Splunk HEC endpoint is configured with a certificate generated by the Puppet CA, set this parameter to false to allow metrics to successfully send.

fips_verify_peer

Boolean: In FIPS mode, the HTTP Client will attempt peer verfication by default. When utilizing a self-signed certificate set this parameter to false to allow metrics to successfully send.

NOTE

To set up a testing environment with FIPS enabled, run the following command: PROVISION_LIST=fips_acceptance bundle exec rake acceptance:setup

Advanced Topics

Known Issues

Breaking Changes

Release Process

This module is hooked up with an automatic release process using Github actions. To provoke a release simply check the module out locally, tag with the new release version, then github will promote the build to the forge.

Full process to prepare for a release:

Update metadata.json to reflect new module release version (0.8.1). Run bundle exec rake changelog to update the CHANGELOG automatically. Submit PR for changes.

Create Tag on target version:

git tag -a v0.8.1 -m "0.8.1 Feature Release"
git push upstream --tags

Authors

P.I.E. Team

P. uppet\ I. ntegrations\ E. ngineering

team-pie@puppet.com