Search Audit Index

The Search Audit Index provides event logs on search usage and activities for your account. The index allows you to monitor and audit the search queries being run within your account, the types of queries, the users running them, and more.

To visualize your Search Audit Index data, download the Enterprise Search Audit App, which provides pre-built dashboards and reports to help you analyze your current search use and identify areas for improvement.

Availability

This feature is available in the following account plans.

Account Type	Account Level
Cloud Flex	Trial, Enterprise
Cloud Flex Credits	Trial, Enterprise Operations, Enterprise Security, Enterprise Suite

Enable the Search Audit Index

The Search Audit Index must be enabled by an administrator.

To enable the Search Audit Index:

1. Classic UI. In the main Sumo Logic menu, select Administration > Security > Policies.
   New UI. In the top menu select Administration, and then under Account Security Settings select Policies. You can also click the Go To... menu at the top of the screen and select Policies.
2. Under Sumo Logic Auditing, select Enable Search Audit Record.
   

note

• Users with a role that grants the Manage audit data feed capability are allowed to enable the Search Audit Index.
• Enabling the index will not count towards your data volume quota.
• Logging to the index begins when the index is enabled.

Query the Search Audit Index

1. In the Search page, enter the query _index=sumologic_search_usage_per_query.
2. Choose the time range for the audit events you’d like to review.
3. Click Start to run the search. Results return in the Messages tab.

note

Querying the index returns results only if the index is enabled.

Index retention period

By default, the retention period of the Search Audit index is the same as the retention period of your Default partition. You can change the retention period by editing the partition that contains the index, sumologic_search_usage_per_query. For more information, see Edit a Partition.

Log Search Audit Index message fields

The following table provides details on the fields returned by the index:

Field	Description
time	The time when the audit log was generated.
analytics_tier	The data tier associated with the audit message. Learn more about Data Tiers.
content_identifier	The ID of the content item that triggered the search query.
content_name	The name of the content item that triggered the search query.
data_retreived_bytes	Amount of data retrieved by the search query. This represents the approximate size of messages that match the source expression of the query and are retrieved from scanning.
data_scanned_bytes	Amount of data scanned by the search query. This value is an approximation, as the scanned message bytes are captured at intermittent time intervals and then averaged over the query time range. It is important to note that this value may be less than the retrieved bytes in some cases due to the approximation. Additionally, if a query contains a timecompare or subquery operator, the data_scanned_byte attribute in the audit log will include the sum of both the parent and child queries.
execution_duration_ms	Time taken to complete the search.
is_aggregate	The boolean variable that indicates if the corresponding search query was an aggregate query. The aggregate operator’s list can be found in Group or Aggregate Operators.
query	The query text string run by the user.
query_end_time	The end time in the time range specified as part of the query time parameter, in milliseconds since epoch.
query_start_time	The start time in the time range specified as part of the query time parameter, in milliseconds since epoch.
query_type	Identifies the type of query run within the account such as API, UI, Scheduled Views, etc. The values and their detailed description are provided in Query type field values.
remote_ip	The remote IP of the source from where the query originated.
retrieved_message_count	The number of messages returned by the search result. This represents the approximate count of messages that match the source expression of the query and are retrieved from scanning.
scanned_bytes_breakdown	Amount of data scanned by the search query broken down by type.
scanned_message_count	The number of messages scanned by the search. This is an approximation, as scanned_message_count is captured at intermittent time intervals and averaged over a query time range. (May be less than retrieved_message_count in some cases due to this approximation.)
scanned_partition_count	The number of partitions scanned by the search. This is an approximation as scanned message bytes are captured at intermittent time intervals and averaged over a query time range. (May be less than retrieved bytes in some cases due to this approximation.)
session_id	An identifier for every search run within the account. This is the same SESSION number displayed in the UI in the search tab.
status_message	Gives the status of the search. The values include: Finished successfully, Query failed, and Query canceled.
user_name	The email of the user that ran the search.

Query type field values 

The table below shows the possible values for the query_type field.

query_type value	Description
Alerts	Search queries run by users accessing the Alert Response Page.
Search API	Search queries run by users using the Search Job API only.
Interactive Search	Search queries run from the Search tab in the UI only.
Interactive Dashboard	Search queries run from dashboards in the UI only.
Scheduled Search	Scheduled search queries run as per the frequency specified by users in the org.
View Maintenance	Scheduled View queries run on behalf of the users in the org.
Sumo Internal	The Internal searches Sumo Logic runs in the background that are critical in providing other services (for example, autocomplete, scheduled view optimization, etc.).
Auto Refresh Dashboard	Search queries used to power auto refresh dashboard panels.
Monitor	Queries associated with monitors.
Span Analytics	Queries run for filtering and aggregating trace data based on span attributes to understand application services performance. Queries can be built using input fields, with filters and visualized results available.