Permissions for alert suppression in machine learning rules (#5819)

* Update ESS docs: ML rule req for alert suppression

* Update serverless docs, and align with ESS

(cherry picked from commit 632deff)

# Conflicts:
#	docs/serverless/advanced-entity-analytics/ml-requirements.mdx
#	docs/serverless/alerts/alert-suppression.mdx

docs/advanced-entity-analytics/ml-req.asciidoc

@@ -7,6 +7,10 @@ To run and create {ml} jobs and rules, you need all of these:
 * There must be at least one {ml} node in your cluster
 * The `machine_learning_admin` user role
 
+Additionally, to configure <<alert-suppression,alert suppression>> for {ml} rules, your role needs the following {kibana-ref}/kibana-role-management.html#adding_index_privileges[index privilege]:
+
+* `read` permission for the `.ml-anomalies-*` index
+
 For more information, go to {ml-docs}/setup.html[Set up {ml-features}].
 
 [IMPORTANT]

docs/detections/alert-suppression.asciidoc

@@ -4,7 +4,9 @@
 .Requirements and notices
 [sidebar]
 --
-Alert suppression requires a https://www.elastic.co/pricing[Platinum or higher subscription].
+* Alert suppression requires a https://www.elastic.co/pricing[Platinum or higher subscription].
+
+* {ml-cap} rules have <<ml-requirements,additional requirements>> for alert suppression.
 
 preview::["Alert suppression is in technical preview for threshold, indicator match, event correlation, and new terms rules. The functionality may be changed or removed in a future release. Elastic will work to fix any issues, but features in technical preview are not subject to the support SLA of official GA features."]
 --
@@ -17,7 +19,7 @@ Alert suppression allows you to reduce the number of repeated or duplicate detec
 * <<create-eql-rule,Event correlation>> (non-sequence queries only)
 * <<create-new-terms-rule,New terms>>  
 * <<create-esql-rule,{esql}>>
-* <<create-ml-rule,{ml-app}>>
+* <<create-ml-rule,{ml-cap}>>
 
 Normally, when a rule meets its criteria repeatedly, it creates multiple alerts, one for each time the rule's criteria are met. When alert suppression is configured, duplicate qualifying events are grouped, and only one alert is created for each group. Depending on the rule type, you can configure alert suppression to create alerts each time the rule runs, or once within a specified time window. You can also specify multiple fields to group events by unique combinations of values.
 

docs/serverless/advanced-entity-analytics/ml-requirements.mdx

@@ -0,0 +1,30 @@
+---
+slug: /serverless/security/ml-requirements
+title: ((ml-cap)) job and rule requirements
+description: Requirements for using ((ml)) jobs and rules.
+tags: [ 'serverless', 'security', 'reference', 'manage' ]
+---
+
+<DocBadge template="technical preview" />
+<div id="ml-requirements"></div>
+
+To run and create ((ml)) jobs and rules, you need the appropriate <DocLink slug="/serverless/general/assign-user-roles">user role</DocLink>.
+
+Additionally, for <DocLink slug="/serverless/custom-roles">custom roles</DocLink>, to configure <DocLink slug="/serverless/security/alert-suppression">alert suppression</DocLink> for ((ml)) rules, your role needs the following index privilege:
+
+* `read` permission for the `.ml-anomalies-*` index
+
+For more information, go to [Set up ((ml-features))](((ml-docs))/setup.html).
+
+<DocCallOut title="Important" color="warning">
+
+Some roles give
+access to the results of _all_ ((anomaly-jobs)), irrespective of whether the user
+has access to the source indices. Likewise, a user who has full or read-only
+access to ((ml-features)) within a given ((kib)) space can view the results of _all_
+((anomaly-jobs)) that are visible in that space. You must carefully consider who
+is given these roles and feature privileges; ((anomaly-job)) results may propagate
+field values that contain sensitive information from the source indices to the
+results.
+
+</DocCallOut>

docs/serverless/alerts/alert-suppression.mdx

@@ -0,0 +1,117 @@
+---
+slug: /serverless/security/alert-suppression
+title: Suppress detection alerts
+description: Reduce noise from rules that create repeated or duplicate alerts.
+tags: [ 'serverless', 'security', 'how-to' ]
+status: in review
+---
+
+<DocBadge template="technical preview" />
+<div id="alert-suppression"></div>
+
+<DocCallOut color="warning" title="Requirements and notice">
+    - ((ml-cap)) rules have <DocLink slug="/serverless/security/ml-requirements">additional requirements</DocLink> for alert suppression.
+
+    - Alert suppression is in technical preview for threshold, indicator match, event correlation, and new terms rules. The functionality may be changed or removed in a future release. Elastic will work to fix any issues, but features in technical preview are not subject to the support SLA of official GA features.
+</DocCallOut>
+
+Alert suppression allows you to reduce the number of repeated or duplicate detection alerts created by these detection rule types:
+
+* <DocLink slug="/serverless/security/rules-create" section="create-custom-rule" text="Custom query" />
+* <DocLink slug="/serverless/security/rules-create" section="create-threshold-rule" text="Threshold" />
+* <DocLink slug="/serverless/security/rules-create" section="create-indicator-rule" text="Indicator match" />
+* <DocLink slug="/serverless/security/rules-create" section="create-eql-rule" text="Event correlation" /> (non-sequence queries only)
+* <DocLink slug="/serverless/security/rules-create" section="create-new-terms-rule" text="New terms" />
+* <DocLink slug="/serverless/security/rules-create" section="create-esql-rule" text="ES|QL" />
+* <DocLink slug="/serverless/security/rules-create" section="create-ml-rule" text="Machine learning" />
+
+Normally, when a rule meets its criteria repeatedly, it creates multiple alerts, one for each time the rule's criteria are met. When alert suppression is configured, duplicate qualifying events are grouped, and only one alert is created for each group. Depending on the rule type, you can configure alert suppression to create alerts each time the rule runs, or once within a specified time window. You can also specify multiple fields to group events by unique combinations of values.
+
+The ((security-app)) displays several indicators in the Alerts table and the alert details flyout when a detection alert is created with alert suppression enabled. You can view the original events associated with suppressed alerts by investigating the alert in Timeline.
+
+<DocCallOut title="Note">
+Alert suppression is not available for Elastic prebuilt rules. However, if you want to suppress alerts for a prebuilt rule, you can duplicate it, then configure alert suppression on the duplicated rule.
+</DocCallOut>
+
+## Configure alert suppression
+
+You can configure alert suppression when you create or edit a supported rule type. Refer to documentation for creating <DocLink slug="/serverless/security/rules-create" section="create-custom-rule" text="custom query"/>, <DocLink slug="/serverless/security/rules-create" section="create-threshold-rule" text="threshold"/>, <DocLink slug="/serverless/security/rules-create" section="create-indicator-rule" text="indicator match"/>, <DocLink slug="/serverless/security/rules-create" section="create-eql-rule" text="event correlation"/>, <DocLink slug="/serverless/security/rules-create" section="create-new-terms-rule" text="new terms" />, <DocLink slug="/serverless/security/rules-create" section="create-esql-rule" text="ES|QL" />, or <DocLink slug="/serverless/security/rules-create" section="create-ml-rule" text="machine learning" /> rules for detailed instructions.
+
+1. When configuring the rule type (the **Define rule** step for a new rule, or the **Definition** tab for an existing rule), specify how you want to group events for alert suppression:
+
+    * **Custom query rule, indicator match, threshold, event correlation (non-sequence queries only), new terms, ((esql)), or ((ml)) rules:** In **Suppress alerts by**, enter 1-3 field names to group events by the fields' values.  
+    * **Threshold rule:** In **Group by**, enter up to 3 field names to group events by the fields' values, or leave the setting empty to group all qualifying events together. 
+
+    <DocCallOut title="Note">  
+    If you specify a field with multiple values, alerts with that field are handled as follows:
+
+    * **Custom query or threshold rules:** Alerts are grouped by each unique value. For example, if you suppress alerts by `destination.ip` of `[127.0.0.1, 127.0.0.2, 127.0.0.3]`, alerts will be suppressed separately for each value of `127.0.0.1`, `127.0.0.2`, and `127.0.0.3`. 
+    * **Indicator match, event correlation (non-sequence queries only), new terms, ((esql)), or ((ml)) rules:** Alerts with the specified field name and identical array values are grouped together. For example, if you suppress alerts by `destination.ip` of `[127.0.0.1, 127.0.0.2, 127.0.0.3]`, alerts with the entire array are grouped and only one alert is created for the group. 
+
+    </DocCallOut>
+
+1. If available, select how often to create alerts for duplicate events:
+
+     <DocCallOut title="NOTE">
+     Both options are available for custom query, indicator match, event correlation, new terms, ((esql)), and ((ml)) rules. Threshold rules only have the **Per time period** option. 
+     </DocCallOut>
+
+    * **Per rule execution**: Create an alert each time the rule runs and an event meets its criteria. 
+    * **Per time period**: Create one alert for all qualifying events that occur within a specified time window, beginning from when an event first meets the rule criteria and creates the alert. 
+
+        For example, if a rule runs every 5 minutes but you don't need alerts that frequently, you can set the suppression time period to a longer time, such as 1 hour. If the rule meets its criteria, it creates an alert at that time, and for the next hour, it'll suppress any subsequent qualifying events.
+
+        <DocImage size="l" url="../images/alert-suppression/-detections-alert-suppression-options.png" alt="Alert suppression options" />
+
+1. Under **If a suppression field is missing**, choose how to handle events with missing suppression fields (events in which one or more of the **Suppress alerts by** fields don't exist):
+
+     <DocCallOut title="NOTE">
+     These options are not available for threshold rules. 
+     </DocCallOut>
+
+    * **Suppress and group alerts for events with missing fields**: Create one alert for each group of events with missing fields. Missing fields get a `null` value, which is used to group and suppress alerts. 
+    * **Do not suppress alerts for events with missing fields**: Create a separate alert for each matching event. This basically falls back to normal alert creation for events with missing suppression fields.
+
+1. Configure other rule settings, then save and enable the rule.
+
+<DocCallOut title="Tips">
+* Use the **Rule preview** before saving the rule to visualize how alert suppression will affect the alerts created, based on historical data.
+* If a rule times out while suppression is turned on, try shortening the rule's <DocLink slug="/serverless/security/rules-create" section="rule-schedule" text="look-back"/> time or turn off suppression to improve the rule's performance.
+</DocCallOut>
+
+## Confirm suppressed alerts
+
+The ((security-app)) displays several indicators of whether a detection alert was created with alert suppression enabled, and how many duplicate alerts were suppressed.
+
+<DocCallOut title="Important" color="warning">
+After an alert is moved to the `Closed` status, it will no longer suppress new alerts. To prevent interruptions or unexpected changes in suppression, avoid closing alerts before the suppression interval ends.
+</DocCallOut>
+
+* **Alerts** table — Icon in the **Rule** column. Hover to display the number of suppressed alerts:
+
+    <DocImage size="xl" url="../images/alert-suppression/-detections-suppressed-alerts-table.png" alt="Suppressed alerts icon and tooltip in Alerts table" />
+
+* **Alerts** table — Column for suppressed alerts count. Select **Fields** to open the fields browser, then add `kibana.alert.suppression.docs_count` to the table.
+
+    <DocImage size="xl" url="../images/alert-suppression/-detections-suppressed-alerts-table-column.png" alt="Suppressed alerts count field column in Alerts table" />
+
+* Alert details flyout — **Insights** section:
+
+    <DocImage size="xl" url="../images/alert-suppression/-detections-suppressed-alerts-details.png" alt="Suppressed alerts Insights section in alert details flyout" />
+
+## Investigate events for suppressed alerts
+
+With alert suppression, detection alerts aren't created for the grouped source events, but you can still retrieve the events for further analysis or investigation. Do one of the following to open Timeline with the original events associated with both the created alert and the suppressed alerts:
+
+* **Alerts** table — Select **Investigate in timeline** in the **Actions** column.
+
+    <DocImage size="m" url="../images/alert-suppression/-detections-timeline-button.png" alt="Investigate in timeline button"/>
+
+* Alert details flyout — Select **Take action** → **Investigate in timeline**.
+
+## Alert suppression limit by rule type
+
+Some rule types have a maximum number of alerts that can be suppressed (custom query rules don't have a suppression limit):
+
+* **Threshold, event correlation (non-sequence queries only, ((esql)), and ((ml)):** The maximum number is the value you choose for the rule's **Max alerts per run** <DocLink slug="/serverless/security/rules-create" section="rule-ui-advanced-params">advanced setting</DocLink>, which is `100` by default.
+* **Indicator match and new terms:** The maximum number is five times the value you choose for the the rule's **Max alerts per run** <DocLink slug="/serverless/security/rules-create" section="rule-ui-advanced-params">advanced setting</DocLink>. The default value is `100`, which means the default maximum limit for indicator match rules and new terms rules is `500`.