Help Center/ Data Lake Insight/ API Reference/ Permissions and Supported Actions/ Introduction
Updated on 2026-01-08 GMT+08:00
Online Debugging
CLI Examples
View PDF
Share

Introduction

You can use Identity and Access Management (IAM) for fine-grained permissions management of your DLI. If your account does not require individual IAM users, you may skip over this section.

IAM lets you manage access to your resources by controlling access for IAM users, user groups, IAM agencies, or trust agencies. IAM supports role/policy-based authorization and identity policy-based authorization.

The following table describes the differences between these two authorization models.

Table 1 Differences between role/policy-based authorization and identity policy-based authorization

Authorization

Core Relationship

Permissions

Authorization Method

Description

Role/Policy-based authorization

User-permission-authorization scope
  • System-defined role
  • System-defined policies
  • Custom policies

Assigning roles or policies to principals

To authorize a user, you need to add it to a user group first and then specify the scope of authorization. It is hard to provide fine-grained permissions control using authorization by user groups and a limited number of condition keys. This method is suitable for small- and medium-sized enterprises.

Identity policy-based authorization

User-policy
  • System-defined identity policies
  • Custom identity policies
  • Assigning identity policies to principals
  • Attaching identity policies to principals

You can authorize a user by attaching an identity policy to it. User-specific authorization and a variety of key conditions allow for more fine-grained permissions control. However, this model can be hard to set up. It requires a certain amount of expertise and is suitable for medium- and large-sized enterprises.

For example, if you need to grant an IAM user permission to create DLI queues in region A and use OBS in region B:

  • With role/policy-based authorization, the administrator needs to create two custom policies and grant both of them to the IAM user to achieve access control.
  • With identity policy-based authorization, the administrator only needs to create one custom identity policy and configure the condition key g:RequestedRegion for the policy, and then attach the policy to the principal or grant the principal the access permissions. Identity policy-based authorization is more flexible than role/policy-based authorization.

Policies and actions in the two authorization models are not interoperable. You are advised to use the identity policy-based authorization model.

For more information about IAM, see IAM Service Overview.

If you use IAM users in your account to call an API, the IAM users must be granted the required permissions. The required permissions are determined by the actions supported by the API. Only users with the policies allowing for those actions can call the API successfully.

DLI only supports identity policy-based authorization for OpenAPI, and not for SQL-related permissions. You are advised to use "resource authorization through the DLI console" or "role/policy-based authorization".

For details about DLI SQL permission control, see DLI SQL Permission Control.

If an IAM user wants to create a DLI queue by calling an API, then in the role/policy-based authorization scenario, the user must be granted permissions that allow the dli:queue:createQueue action.

If an IAM user wants to query queue details by calling an API, then in the identity policy-based authorization scenario, the user must be granted permissions that allow the dli:queue:get action.