Managing generative AI features in Amazon CodeCatalyst
Amazon Q Developer in Amazon CodeCatalyst includes generative AI features that can help users in projects
in your space develop software faster. Developers frequently have more tasks to do than
time to accomplish them. CodeCatalyst integrates with Amazon Q Developer to provide features that can help
team members accomplish their tasks more quickly and increase the time they have to focus on
the most important parts of their work. These features are only available to users if you
enable generative AI features for your space. If you choose to allow access to these
features, users can access and use these features to help them accomplish their work more
quickly. When these features are enabled, an individual user's usage of and quotas for using
Amazon Q features depends on the user's subscription to Amazon Q Developer. For more information, see
Amazon Q Developer Pricing
Important
Generative AI features are only available in the US West (Oregon) Region.
The generative AI features available for your space provide the following functionality:
-
Assign issues to Amazon Q feature with Amazon Q Developer Agent for software development: Users with the Project administrator or Contributor role in a project can assign issues to Amazon Q. Once assigned, Amazon Q will analyze an issue based on its title and its description, review the code in the specified repository, and attempt to create a draft solution for users to evaluate. Users can assign issues to address problems or feature requests in code. They can also use this feature to create or update workflows for a project. This feature includes interactive commenting between users and Amazon Q in not only the issue, but in any tasks or pull requests it creates. Users can choose to have Amazon Q create one revision of any pull request it creates based on user comments left for Amazon Q.
-
Write description for me: This feature helps users creating pull requests create detailed descriptions of the code changes contained in the pull request by comparing the code in the source and destination branches and evaluating the impact of the differences on the overall application.
-
Create comment summary: This feature helps users reviewing pull requests understand the overall direction of the comments left on the code changes by other reviewers by summarizing the requests and sentiments expressed in all comments in the overview of the pull request.
Note
Powered by Amazon Bedrock: AWS implements automated abuse detection. Because the Write description for me, Create content summary, and Assign issues to Amazon Q feature with Amazon Q Developer Agent for software development are built on Amazon Bedrock, users can take full advantage of the controls implemented in Amazon Bedrock to enforce safety, security, and the responsible use of artificial intelligence (AI).
The generative AI features in CodeCatalyst are subject to quotas. For more information, see
Amazon Q Developer Pricing
Enabling or disabling generative AI features for a space
Before users can use any of the generative AI features in CodeCatalyst, you must first choose to enable them for your space. You cannot enable these features for specific projects. If you enable them for the space, all projects in the space will have access to them.
The generative AI features in CodeCatalyst are useful and powerful tools for users in your space. Like all powerful tools, they perform best when well understood by users and administrators. You should be aware of the following considerations before you decide to enable generative AI features for your space.
-
Generative AI features cannot make arbitrary changes to your project. While Amazon Q can create pull requests with code changes, it cannot merge a pull request. A user must approve the pull request and merge the changes. However, if you have one or more workflows configured to run based on branch events, those workflows will run immediately when the pull request is created. Additionally, once code is merged, the changes are part of the source repository branch where the pull request was merged. Just like any other merged pull request, any workflows configured to build and deploy pushes to the destination branch will start a run of the changes changes contained in a pull request created by Amazon Q.
-
Generative AI features might not correctly interpret user statements and existing code structures. User intent can be misinterpreted by generative AI features, which can result in code generation that does not do what the original requestor expected. The approach Amazon Q creates is based on the title and description of an issue, as well as its analysis of the code in the source repository specified in the issue. If the code in the source repository contains errors, Amazon Q might repeat these errors if it interprets them as part of the desired structure and patterns in the code. As a best practice, encourage users in a space with generative AI features enabled to thoroughly review any code in pull requests created by Amazon Q to ensure that the code both does what the issue title and description suggested, and that the code itself is well-formed and functional. Like any other pull request, users can create a Dev Environment and make changes and fixes to the code in the branch created by Amazon Q. Similarly review a revision created by Amazon Q to help ensure that the changes correctly address any issues raised in comments.
-
Workflows created by Amazon Q might deploy code to production resources immediately once a pull request is merged. By default, CodeCatalyst does not allow Amazon Q to create or update workflows. However, you can choose to have Amazon Q do so, and it will try to create or update a workflow according to the information in the issue. If you merge a pull request with workflow changes, that workflow will automatically run when its conditions are met. Additionally, like any pull request, pull requests created by Amazon Q will start a workflow run when the specified conditions are met, so if you have a workflow configured to run on branch events, a pull request with workflow changes might start a run of those workflows.
You must have the Space administrator role in CodeCatalyst to manage generative feature access for a space.
Important
Generative AI features are only available in the US West (Oregon) Region.
To enable or disable generative AI features for a space
Open the CodeCatalyst console at https://codecatalyst.aws/
. -
Navigate to your CodeCatalyst space. Choose Settings, and then choose Generative AI.
The Generative AI page displays.
-
To enable generative AI features for all projects in your space, choose the Projects in this space can access generative AI features toggle and make sure it is in the on position. Users in projects will immediately have access to the generative AI features available to their roles in those projects.
-
To disable generative AI features for all projects in your space, choose the Projects in this space can access generative AI features toggle and make sure it is in the off position.
Warning
Disabling the generative AI features will stop work on all issues and summaries in all projects. The work cannot be restarted even if you immediately re-enable the generative AI features.
Viewing usage of generative AI features in a space
Although generative AI features are enabled for a space, the individual usage of those features is captured at the individual user level. You can view information about your individual usage of generative AI features by going to your My settings page.
User usage and quotas for using generative AI features in CodeCatalyst is based on that user's subscription
to Amazon Q Developer. For more information, see Amazon Q Developer Pricing
To view your generative AI feature usage
Open the CodeCatalyst console at https://codecatalyst.aws/
. -
Choose your provfile gravitar, and then choose My settings.
Review the information in Generative AI feature usage to determine current usage levels and whether your individual user account is close to exceeding any quotas.
