Loading

Triage and manage attacks

The Attacks page is the triage hub where you view, filter, assign, tag, and act on all attacks in your environment, alongside their correlated alerts. It pairs with the Attack Discovery page, which is where you manually generate new discoveries using large language models (LLMs):

  • Go to Attack Discovery to run LLM analysis on demand and create new attack discoveries.
  • Go to Attacks for day-to-day triage of all attacks (manual and scheduled), and to manage their investigation lifecycle.

You can also schedule recurring Attack Discovery runs from either page; schedules created on one page appear on the other.

The Attacks page requires the same privileges as Attack Discovery. Refer to Role-based access control (RBAC) for Attack Discovery for details.

Important

To access the Attacks page, you must turn on the Enable alerts and attacks alignment setting under Security Solution in Advanced Settings.

At the top of the Attacks page, you can find overview visualizations and tables. The Summary tab shows the total number of attacks detected and attack volume over time, while the Trends, Count, and Treemap tabs all describe alerts associated with these attacks.

Overview of the Attacks page showing the Summary tab

You can schedule Attack Discovery runs directly from the Attacks page. The scheduling flow is the same as on the Attack Discovery page, and schedules you create on either page appear on both. For step-by-step instructions, refer to Schedule discoveries.

The Attacks table appears under the summary section and lists individual attacks. You can expand an attack to view details including which entities were involved and which steps of the attack chain were performed.

Use the controls at the top of the Attacks table to narrow results:

Filter method Description
KQL search Enter queries in the search bar. Autocomplete includes fields from both attacks and alerts.
Date/time picker Set a specific time range.
Status filter Filter by Open, Acknowledged, or Closed.
Assignees filter Click Filter by assignees to show only attacks or alerts assigned to specific users.
Sort Use the Sort by menu to sort by Most recent, Least recent, Most alerts, or Least alerts.

Click the View options ( ) menu to access the following toggles:

Show attacks only: This toggle is enabled by default. It hides standalone alerts (alerts that don't belong to any attack) so you can focus on correlated attack groups. Disable it to see all alerts, including those not linked to any attack.

Note

When Show attacks only is disabled, standalone alerts appear in a group labeled - (dash). This group acts as a bucket for all alerts that aren't linked to an attack.

Show anonymized values: When enabled, replaces attack titles and summaries with anonymized placeholder values. If you're searching the page for specific entities like hostnames or IP addresses, make sure to turn this off.

The Attacks page uses a single data view that combines both the attacks index and the alerts index. This enables powerful cross-document filtering, but it also means that filters apply to both document types simultaneously.

Access actions from the Take actions menu on an attack's row in the Attacks table.

Note

When you change an attack's status, assign or unassign it, or apply attack tags, a modal appears that lets you apply the action to the attack only, or to both the attack and its linked alerts.

Action Description
Change status Mark as acknowledged or closed
Run workflow Run an Elastic workflow for on-demand response or investigation
Assign or unassign attack Assign analysts to investigate
Apply attack tags Categorize attacks for filtering
Investigate in timeline Open the attack in Timeline for analysis
Add to case Attach the attack to a new or existing case
View in AI Chat Continue investigating with an AI agent

Attack statuses track investigation progress:

Status Meaning
Open Needs investigation (default)
Acknowledged Under active investigation
Closed Resolved

To change an attack's status, click Take actions on the attack row, then select Mark as acknowledged or Mark as closed.

You can run an Elastic workflow directly from an attack to trigger an on-demand response or investigation. To use this feature, make sure you meet the workflows prerequisites.

To run a workflow, click Take actions on the attack row, then click Run workflow. Use the search bar to select a workflow, then click Run workflow.

Note

You can select only enabled workflows.

Assign analysts to attacks they should investigate.

Task How to do it
Assign an attack Take actions > Assign attack > select users
Unassign an attack Take actions > Unassign attack
Important

Users are not notified when assigned or unassigned.

Tags help organize attacks into filterable categories.

To tag an attack, click Take actions, then select Apply attack tags.

To open an attack in Timeline, click Take actions on the attack row, then select Investigate in timeline.

Note

Investigating an attack in Timeline includes all alerts that were originally correlated when the attack was created. It doesn't reflect your current page filters or selected time range.

To add an attack to a case, click Take actions, then select Add to new case or Add to existing case.

To continue investigating an attack with an AI agent, click Take actions, then select View in AI Chat. You can ask follow-up questions about the attack or its associated alerts.