All Products
Search
Document Center

AgentLoop:AgentSpace

Last Updated:Aug 10, 2026

An AgentSpace is the basic unit for organizing resources in AgentLoop, providing isolation, access control, and billing attribution. Create an AgentSpace before you build agents so that datasets, evaluations, memory stores, and observability data are grouped under a single, permissioned boundary.

AgentSpace contents

An AgentSpace groups two categories of resources:

  • AgentLoop-owned assets — Datasets, evaluation tasks, memory stores, and other assets created directly in AgentLoop.

  • Associated Alibaba Cloud resources — External resources in Cloud Monitor 2.0, Simple Log Service (SLS), and MSE (Microservices Engine) AI Governance Center that host observability data, logs, and agent assets.

Associated resources

Each AgentSpace binds to the following Alibaba Cloud resources. The table shows which resources you provide and which the system creates for you during space creation.

Resource

Purpose

Provisioning

Cloud Monitor 2.0 workspace

Hosts agent call-chain and model monitoring data.

You choose to auto-create a new workspace or reuse an existing one.

SLS project

Agent Trace,hosts agent audit logs, evaluation logs, and other log data.

Created automatically by the system.

MSE AI Governance Center namespace

Hosts Prompts, Skills, and other agent assets.

Created automatically by the system.

Prerequisites

Before you create an AgentSpace, prepare the following. The distinction between activating a service and creating a resource instance is important: you must activate the services yourself, but AgentLoop provisions the individual SLS project and MSE namespace on your behalf during space creation.

  • An Alibaba Cloud account with real-name verification completed.

  • Cloud Monitor 2.0, Simple Log Service (SLS), and MSE (Microservices Engine) AI Governance Center activated under that account.

  • The AliyunAgentLoopFullAccess permission policy granted to the RAM user that will perform the creation.

AgentSpace creation

  1. Log on to the AgentLoop console.

  2. On the console homepage, click Create Space.

  3. In the Create AgentSpace dialog box, configure the following parameters.

    Parameter

    Description

    Region

    Select the region for the AgentSpace. Cannot be changed after creation.

    AgentSpace

    The space name. Enter 3 to 63 characters using lowercase letters, digits, hyphens (-), or underscores (_). The name must be globally unique across all users and cannot be changed after creation.

    Description

    A brief description of the space purpose.

    Cloud Monitor 2.0 Workspace

    Choose one of the following options. The binding cannot be changed after creation.
    Auto-create — The system creates a new workspace named agentloop-<32-character code>.
    Select existing — Reuse an existing Cloud Monitor 2.0 workspace in the same region under the current account, which is convenient for maintaining unified monitoring views with existing applications.

    Agent Trace

    Checked by default. When enabled, the system processes incoming Traces into ATIF-standard Agent Traces, allowing you to view the complete execution chain and facilitate subsequent data processing. If not currently needed, you can uncheck this option.

  4. Click Create Space. The system creates the AgentSpace, provisions the SLS project and MSE AI Governance Center namespace, and completes resource binding. You can then find the new space under System Management > Space Management.And set the Agent Trace collection status according to the selected configuration.

Limits

  • Region binding — The region is fixed at creation time and cannot be changed afterward.

  • Name uniqueness — The AgentSpace name must be globally unique across all Alibaba Cloud users. If the name is already taken, choose a different one.

  • Workspace binding — The associated Cloud Monitor 2.0 workspace cannot be changed after creation.

What to do next

After the AgentSpace is created, go to System Management > Space Management to:

  • View space details, including the associated Cloud Monitor 2.0 workspace, SLS project, and MSE AI Governance Center namespace.

  • Modify configurations that remain editable, such as the description.

  • Delete the space when it is no longer needed.