Skip to content

OWL

OWL is the unified capability entry point provided by TrueWatch for tool-oriented and intelligent scenarios, supporting two access modes: OWL CLI and OWL MCP Server.

Through OWL, users can uniformly call capabilities such as metrics, logs, events, monitors, dashboards, APM, RUM, and infrastructure in TrueWatch.

OWL does not replace the TrueWatch console or Open API. Instead, it provides a wrapper entry point that is more suitable for tool invocation, automated orchestration, and AI Agent usage, built on top of existing capabilities.

How It Works

flowchart LR
  user["User / Script"] --> cli["OWL CLI"]
  agent["AI Agent / MCP Client"] --> mcp["OWL MCP Server"]
  cli --> owl["OWL Tool Catalog and Execution Service"]
  mcp --> owl
  owl --> platform["TrueWatch Platform Capabilities"]

OWL CLI and OWL MCP Server access the same set of platform capabilities but are designed for different access modes and use cases.

Access Mode Use Cases Core Capabilities
OWL CLI Local command line, scripts, and automation workflows Sync tool catalog, view tool descriptions, execute tools, export Agent Schema
OWL MCP Server AI Agent, MCP client integration Provide standardized tool service interfaces based on Model Context Protocol

Use Cases

  • Query metrics, logs, events, monitors, dashboards, and other information in the terminal
  • Quickly verify observability capabilities in customer workspaces during implementation and delivery
  • Integrate TrueWatch capabilities into AI clients that support MCP
  • Enable agents to call TrueWatch query, analysis, and management tools using natural language
  • Leverage capabilities such as the same-org workspace list, dedicated cross-workspace Trace query tool (owl.data.same_org.trace.query), SLO list (owl.slo.list), and documentation search (mdsearch) to conduct cross-workspace troubleshooting and queries

Usage Limitations

  • OWL's access permissions are tied to the API Key used. Resources that the API Key does not have permission to access are also inaccessible to OWL.
  • Write-type tools will modify resources in TrueWatch, such as creating dashboards, updating monitors, appending comments, etc.
  • When integrating with AI Agents or MCP clients, it is recommended to configure manual confirmation or approval mechanisms for write operations to avoid unintended changes.

Quick Start

Scenario 1: Command Line Operations

Suitable for developers, operations personnel, and implementation and delivery personnel to manually execute queries and management tasks.

Scenario 2: AI Agent Integration

Suitable for integrating platform capabilities into large model applications to achieve natural language interaction.