Bits AI Dev Agent is a generative AI coding assistant that uses observability data from Datadog to automatically diagnose and fix issues in your code. It integrates with GitHub to create production-ready pull requests, iterates on fixes using CI logs and developer feedback, and draws on multiple Datadog products to generate contextual fixes.
Supported Datadog products
Bits AI Dev Agent is available for the following Datadog products:
Remediates code vulnerabilities individually or in bulk
Test Optimization
Preview
Provides code fixes for flaky tests and verifies that tests remain stable
Continuous Profiler
Preview
Provides code changes for Automated Analysis insights
Containers
Preview
Provides code changes for Container Recommendations
Note: Enabling Bits AI Dev Agent is product-specific. Even if it’s active for one Datadog product, it must be separately enabled for each additional product you use.
Key capabilities
The following sections detail how Bits AI Dev Agent integrates with Datadog products to generate contextual code fixes.
Pull request assistance
Bits AI Dev Agent integrates with GitHub to create pull requests, respond to comments, update commits, and fix CI failures.
Generates PR titles and descriptions based on your PR template.
Opens PRs as drafts, iterates using CI logs, and marks the PRs as ready for review when checks pass.
Continues iterating in response to chat messages and review feedback.
Note: Comment @Datadog to prompt Bits for updates to the PR. Bits Dev never auto-merges PRs.
Go to Bits AI > Dev Agent > Code sessions to see all Dev Agent code sessions and generated PRs. You can search sessions and filter by service, product source, and status.
Auto-push
Auto-push allows the Dev Agent to create branches, push code, and open PRs when it detects something it can help you with. For example, the Dev Agent can:
Auto-create PRs for high-impact errors (such as 500s or crashes).
Update PRs in response to your comments in GitHub.
Update PRs to address CI failures.
Auto-push only opens PRs and pushes changes; it never merges code. When auto-push is disabled, you must review code in Datadog before it gets pushed.
Auto-push is available for Error Tracking and Test Optimization.
Security considerations
Allowing any AI-based tool to read untrusted data can let attackers trick it into outputting malicious code or other output. In some environments, an attacker could craft errors, traces, or other telemetry containing malicious payloads that the Dev Agent reads. Datadog runs security scanning on the output of the Dev Agent, but it is not foolproof.
Error tracking
In Error Tracking, Bits AI Dev Agent diagnoses and remediates code issues with context and unit-tested fixes:
Determines whether an error can be fixed through code and generates a fix with unit tests.
Provides links within the chat to relevant files and methods for streamlined navigation.
Analyzes errors asynchronously as they arrive.
Marks errors with a Fix available status and enables filtering to surface those issues.
Bits AI Dev Agent fixes flaky tests that are detected through Flaky Test Management in Test Optimization and attempts to verify that tests remain stable.
