[Alerting V2] [DOCS] [M2] Document that activated episodes stay open until manually closed

[nastasha-solomon]

Summary

The Alerting v2 core engine now reacts immediately (via the dispatcher, not waiting for the next rule execution) when a user activates or deactivates an episode via UI or API. Key behavioral decision: once a user activates an episode, the engine keeps it open and ignores automatic recoveries for that episode until the user manually closes it; deactivating lets a new episode open on the next breach.

Why this needs docs: triage-alert-episodes.md (PR #6527) documents 'Activate: Manually moves the episode to active state without waiting to meet the activation threshold' but doesn't mention the critical follow-on behavior that an activated episode is then exempted from automatic recovery and must be manually closed — an omission a user needs to know to avoid assuming the episode will auto-resolve normally.

Resources

• PR #274618 — [Alerting v2] Take into account activated/deactivated episodes in the core engine
• Product issue: [Alerting v2] — code/details from Boom.data are dropped from 4xx/5xx response bodies kibana#274951

Availability

Channel	Details
Stack	v9.5.0
Serverless	Jul 13–Jul 17
Feature status	preview
Feature flag	None — active by default

---

Created with Docs Quest Scanner by @nastasha-solomon

Suggested edits

Triage alert episodes > Available actions

• What the docs say: "Activate: Manually moves the episode to `active` state without waiting to meet the activation threshold." No mention of what happens to the episode's automatic recovery behavior after manual activation.
• What to add: This page (still in open docs-content PR [Alerting V2][Serverless & 9.5][M2] Add experimental alerting features alerts pages #6527) describes the Activate action but doesn't state that once an episode is manually activated, the engine ignores automatic recoveries for it and it stays open until manually deactivated/closed — user intention takes priority over the rule's own recovery detection. Without this, a user could reasonably expect the episode to auto-recover normally like any other active episode. Applies from 9.5.0 (experimental) and in serverless (experimental).

[github-actions]

Elastic Docs AI Menu

Check one box to run an AI workflow for this issue.

• Triage (docs-triage). Status: not started.
• Size the work (docs-size). Status: not started.
• Scope the docs work (docs-issue-scope). Status: not started.

Powered by GitHub Agentic Workflows and docs-actions. For more information, reach out to the docs team.

[nastasha-solomon]

Addressed in 491edf9 (#6527).