Registering and Deploying Micro Frontends (MFE) in DataHub Cloud – DataHub Customer Support Portal

```html

Area: Deployment Issues

Sub-Area: Micro Frontend (MFE) Registration and Configuration

Issue

Organizations that have built and validated a Micro Frontend (MFE) on a local DataHub instance want to deploy it to a DataHub Cloud environment. The process involves two distinct responsibilities: hosting the MFE bundle (customer-managed) and registering the MFE configuration in the managed DataHub Cloud frontend (support-managed). There is no self-service mechanism for customers to register or update MFE configurations directly in DataHub Cloud; all registration requires a support-initiated configuration update and frontend redeploy.

You Might Be Asking

Can I register my MFE myself in DataHub Cloud, or does the DataHub team need to do it?
What configuration do I need to provide to get my MFE loaded in DataHub Cloud?
How does DataHub Cloud load MFE bundles, and what CORS settings are required?
Is MFE support available in DataHub Cloud, and since which version?

Solution

Deploying an MFE in DataHub Cloud is a two-part process split between the customer and the DataHub support team.

Host your MFE bundle (customer responsibility):
Serve your MFE's remoteEntry.js and associated chunk files over HTTPS from a publicly accessible location (e.g., a cloud storage bucket). Configure CORS on your hosting origin to allow requests from your DataHub Cloud instance domain:
```
# Example CORS configuration (conceptual)
# Allow your DataHub Cloud instance origin
Access-Control-Allow-Origin: https://<your-instance>.acryl.io
```
Your MFE's GraphQL calls to /api/v2/graphql are same-origin requests that rely on the logged-in user's session cookie. No additional authentication configuration or secrets are required on your end.

Ensure your MFE exposes a mount(containerElement, {}) function via Module Federation's standard remoteEntry.js contract. The DataHub MFE loader calls mount directly and expects an optional cleanup function returned from it.
Provide your MFE YAML configuration to DataHub Support (support responsibility):
There is currently no self-service way for customers to register an MFE in DataHub Cloud. The DataHub Cloud frontend service reads an mfe.config.yaml file at startup, with the file path specified via the MFE_CONFIG_FILE_PATH environment variable. This configuration is loaded once and cached in memory for the lifetime of the container, so any change requires both a configuration update and a frontend redeploy by the DataHub support team.

Submit a support request and include your MFE YAML configuration block. A minimal example looks like the following:
```
microFrontends:
  - name: <your-mfe-name>
    remoteEntry: https://<your-hosting-domain>/remoteEntry.js
    module: <YourModuleName>
    scope: <yourMfeScope>
```
Replace the placeholder values with your actual MFE details:
- <your-mfe-name> — a unique identifier for your MFE
- https://<your-hosting-domain>/remoteEntry.js — the public HTTPS URL where your bundle is served
- <YourModuleName> — the Module Federation exposed module name
- <yourMfeScope> — the Module Federation scope/container name
DataHub Support performs the configuration update and redeploy:
Once your YAML configuration is received, the DataHub support team will update the mfe.config.yaml for your target Cloud environment and redeploy the datahub-frontend service. After the redeploy completes, your MFE will be loaded and available in your DataHub Cloud instance.

Additional Notes

MFE support in DataHub Cloud has been available since the v0.3.16 release. The requirement for support-driven registration is not a limitation or regression — it is the intended operational model for DataHub Cloud. All infrastructure and environment-level configuration changes are managed by the DataHub team for security and operational consistency; customers do not have direct access to the underlying Kubernetes configuration or environment variables. Because the MFE config is loaded once at container startup and cached in memory, changes can only take effect after a frontend redeploy. Plan accordingly when scheduling MFE updates or additions in production environments.

Issue

You Might Be Asking

Solution

Additional Notes

Related Documentation