Splunk Events

StackState Self-hosted v5.0.x

This page describes StackState version 5.0.

Go to the documentation for the latest StackState release.

Overview

When the Splunk StackPack has been installed in StackState, you can configure the Splunk Events check on StackState Agent V1 to begin collecting Splunk events data.

Events are collected from Splunk by executing Splunk saved searches that are configured in the StackState Agent V1 Splunk Events check. In order to receive Splunk events data in StackState, you will therefore need to add configuration to both Splunk and StackState Agent V1:

• In Splunk, there should be at least one saved search that generates the events data you want to retrieve.

• In StackState Agent V1, a Splunk Events check should be configured to connect to your Splunk instance and execute the relevant Splunk saved searches.

The Splunk Events check on StackState Agent V1 will execute all configured Splunk saved searches periodically. Data will be requested from the last received event timestamp up until now.

Splunk saved search

Fields used

StackState Agent V1 executes the Splunk saved searches configured in the Splunk Events Agent check configuration file and pushes retrieved data to StackState as a telemetry stream. The following fields from the results of a saved search are sent to StackState:

Field	Type	Required?	Description
_time	long	✅	The data collection timestamp, in milliseconds since epoch.
event_type	string	-	Event type, for example server_created.
msg_title	string	-	Message title.
msg_text	string	-	Message text.
source_type_name	string	-	Source type name.
All other fields	-	-	Splunk default fields (docs.splunk.com) other than _time will be filtered out of the result. Any other fields present in the result will be mapped to tags in the format field:value.

Example Splunk query

Example Splunk query

index=monitor alias_hostname=*
| eval status = upper(status)
| search status=CRITICAL OR status=error OR status=warning OR status=OK
| table _time hostname status description

The example Splunk saved search above would result in the following event data in StackState:

Field	Data
_time	Splunk _time field.
event_type	-
msg_title	-
msg_text	-
source_type_name	-
tags	hostname:<hostname> status:<status> description:<description>

Agent check

Configure the Splunk Events check

To enable the Splunk Events integration and begin collecting events data from your Splunk instance, the Splunk Events check must be configured on StackState Agent V1. The check configuration provides all details required for the Agent to connect to your Splunk instance and execute a Splunk saved search.

Example Splunk Events Agent check configuration file: splunk_event/conf.yaml.example (github.com)

To configure the Splunk Events Agent check:

1. Edit the StackState Agent V1 configuration file /etc/sts-agent/conf.d/splunk_events.yaml.

2. Under instances, add details of your Splunk instance:

   • url - The URL of your Splunk instance.

   • authentication - How the Agent should authenticate with your Splunk instance. Choose either token-based (recommended) or basic authentication. For details, see authentication configuration details.

   • tags - Optional. Can be used to apply specific tags to all reported events in StackState.

3. Under saved_searches, add details of each Splunk saved search that the check should execute:

   • name - The name of the Splunk saved search to execute.

     • match - Regex used for selecting Splunk saved search queries. Default "events.*".

     • app - The Splunk app in which the saved searches are located. Default "search".

     • request_timeout_seconds - Default 10.

     • search_max_retry_count - Default 5.

     • search_seconds_between_retries - Default 1.

     • batch_size - Default 1000.

     • initial_history_time_seconds - Default 0.

     • max_restart_history_seconds - Default 86400.

     • max_query_chunk_seconds - Default 3600.

     • unique_key_fields - The fields to use to uniquely identify a record. Default _bkt and _cd.

     • parameters - Used in the Splunk API request. The default parameters provided make sure the Splunk saved search query refreshes. Default force_dispatch: true and dispatch.now: true.

4. More advanced options can be found in the example configuration (github.com).

5. Save the configuration file.

6. Restart StackState Agent V1 to apply the configuration changes.

7. Once the Agent has restarted, wait for the Agent to collect data and send it to StackState.

8. Events retrieved from splunk are available in StackState as a log telemetry stream in the stackstate-generic-events data source. This can be mapped to associated components.

Uniquely identify a record

To prevent sending duplicate events over multiple check runs, received saved search records must be uniquely identified for comparison. By default, a record is identified by the Splunk default fields _bkt and _cd. This behavior can be customized for each saved search by specifying unique_key_fields in the Splunk Events Agent check configuration. Note that the specified unique_key_fields fields are mandatory fields for each record returned by the Splunk saved search.

If it is not possible to uniquely identify a record by a combination of specific fields, the whole record can be used by setting unique_key_fields: [] (an empty list).

Disable the Agent check

To disable the Splunk Events Agent check:

1. Remove or rename the Agent integration configuration file, for example:

    mv conf.d/splunk_event.yaml conf.d/splunk_event.yaml.bak

2. Restart the StackState Agent(s) to apply the configuration changes.

Splunk Events in StackState

Events retrieved from splunk are available in StackState as a log telemetry stream in the stackstate-generic-events data source. This can be mapped to associated components.

See also

• StackState Splunk integration details

• Map telemetry to components

• Example Splunk Events configuration file - splunk_event/conf.yaml.example (github.com)

• Splunk default fields (docs.splunk.com)