--- title: AWS CloudWatch MCP server AI agent description: This AI agent automates AWS CloudWatch alert investigations by analyzing alarm details, querying affected resources, and retrieving CloudWatch metrics and logs across multiple strategies. It provides clear summaries with root cause analysis and actionable recommendations for resolution. locale: en-US canonical_url: https://www.servicenow.com/docs/r/australia/intelligent-experiences/itom-obs-aws-cloudwatch-mcp-server-agent-ai-agent.html release: australia topic_type: reference last_updated: "2026-09-10" reading_time_minutes: 2 breadcrumb: [IT Operations Management AI agents, IT Operations Management, AI agents library, AI assets, Enable AI experiences] --- # AWS CloudWatch MCP server AI agent This AI agent automates AWS CloudWatch alert investigations by analyzing alarm details, querying affected resources, and retrieving CloudWatch metrics and logs across multiple strategies. It provides clear summaries with root cause analysis and actionable recommendations for resolution. ## Workflow 1. Identify what to investigate. The agent uses a provided CloudWatch alarm name or identifies an active alarm associated with a provided entity. 2. Gather context about the alarm and investigate metrics and logs. 3. Return findings, including a summary of alarm history, probable causes, and recommended next steps. For more information about configuring this AI agent, see [Configure observability agents for ServiceNow Otto](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/it-operations-management/configure-integration-agents-for-now-assist.md).
| Field | Description |
|---|---|
| Allow third party to access this AI agent | When enabled, third-party AI agents can use this agent. This value is off \(false\) by default. This setting is defined in the AI Agent configs \[sn\_aia\_agent\_config\] table on the External discoverable field. |
| Allow AI specialists to access this AI agent | When enabled, AI specialists can use this agent. This value is off \(false\) by default. When set to true, more configuration options for tools become available so that an AI specialist can map inputs and response templates to tool outputs. This setting is defined in the AI Agent configs \[sn\_aia\_agent\_config\] table on the Specialist enabled field. |
| Manage long-term memory | When enabled, all previous user interactions are used as context for the LLM. This value is off \(false\) by default. This setting is defined by the **sn\_aia.ltm.enable\_long\_term\_memory** system property. For more information, see [ServiceNow Otto AI agents reference](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/australia/intelligent-experiences/na-aia-reference.md). |
| Tools | - **Model Context Protocol tools** Analyze Metric Cancel Log Insights Query Describe Log Groups Execute Log Insights Query Get Active Alarms Get Alarm History Get Log Insights Query Results Get Metric Metadata Get Recommended Metric Alarms Analyze Log Group Get Metric Data - **Script** Lookup CMDB CI |
| Allowed user roles The specific user roles that can access this AI agent. | evt\_mgmt\_operator |
| Data access roles The specific user identity roles that determine which data the AI agent can access and what actions it can take. | evt\_mgmt\_operator |
| Triggers | Optional. None defined by default. An admin can specify triggers if desired. For more information, see [Add a trigger to an AI agent](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/australia/intelligent-experiences/add-trigger-aia.md). |
| Channels | Configure an assistant for Virtual Agent or ServiceNow Otto panel using [Assistant Designer](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/conversational-interfaces/configure-now-assist-va.md). |
| Used in agentic workflows | [Analyze alert impact agentic workflow](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/it-operations-management/now-assist-itom-agentic-aia.md) |