Plugins

Docs > Internal Developer Portal > Plugins

Join the Preview for Plugins!

Overview

Plugins let developers build, deploy, and manage fully custom applications natively within Datadog using local development workflows and the Datadog CLI.

Plugins introduce a local, code-first development model within Datadog, building on App Builder concepts to support platform engineers and developers who prefer working with code instead of visual editors.

A plugin combines a React-based frontend, serverless backend functions, and Datadog-managed infrastructure, including Datastore, secret management, and telemetry.

Developer workflow

Develop and deploy plugins directly from your local environment using the Datadog CLI.

1. Install the CLI

npm install -g @Datadog/dd-app-cli

2. Scaffold a plugin

dd-app create --agent=claude my-plugin

This command generates the following project structure:

/frontend   → React components
/backend    → Serverless functions (Node.js handlers)
package.json
AGENT.md    → AI agent guidance and structure
config.json → Plugin metadata and UUID

3. Develop locally

dd-app dev

Output:

➜ Local: http://localhost:3000/

This command starts a local development server with live preview. Frontend updates hot-reload automatically, and backend logic runs locally for fast iteration.

4. Deploy to Datadog

dd-app deploy

Authenticate through your browser and publish your plugin:

https://app.datadoghq.com/app-builder/apps/65afc2dc-60de-455b-8d32-64950504fbec

Plugins are automatically instrumented with telemetry and can appear anywhere Datadog supports App Builder apps, such as dashboards, notebooks, and the Internal Developer Portal.

Key features

End-to-end development flow: Build multi-folder React apps locally using CLI commands, AI agents, and Datadog SDKs.
Full-stack capabilities: Combine frontend (React) and backend (serverless) logic using Datadog Actions, including private actions and Lambda-style handlers.
Datadog integrations: Use Datastore, Connections, and RBAC for secure and seamless data and secret management.
Preview and publish: Use dd-app preview for local testing and dd-app publish for one-click deployment to Datadog.
AI assistance: AI agents in AGENT.md provide inline code generation, scaffolding, and refactoring guidance.
Collaboration and GitOps: Integrate with GitHub Sync for pull requests, code reviews, and automated deployments.
Security and observability: All plugins include built-in telemetry, error monitoring, and code security scanning.
Platform embedding: Embed plugins across Datadog surfaces, such as entity pages, self-service actions, homepages, or custom pages.
Consistent UI: Use Druids, Datadog’s UI component library, to ensure a unified design and experience.

Permissions

Plugins inherit the App Builder permissions model:

Plugins execute as the creator, requiring resolve permissions on any connections used.
App-level permissions control who can deploy or modify a plugin.
Connectionless Datadog actions run with the app creator’s credentials.

Architecture

The plugin framework provides an opinionated scaffold optimized for the Datadog ecosystem.

Frontend

Built with React and Druids components.
Automatically instrumented with RUM and error tracking.

Backend

Lightweight serverless functions triggered by Datadog Actions.
Written in TypeScript using the @Datadog/actions SDK.

// backend/getRecentPRs.ts
import * as actions from '@Datadog/actions';
import * as CONNECTIONS from './connections';

export async function getRecentPRs() {
  const results = await actions.github.listPullRequests({
    connectionId: CONNECTIONS.githubConnectionId,
  });
  return results.filter();
}

Frontend usage

// frontend/Root.tsx
import { useBackendQuery } from '@Datadog/react-action-query';

export default function Root() {
  const results = useBackendQuery('getRecentPRs');
  if (results.isLoading) return 'Loading...';
  return <Table data={results.data} />;
}