Code App disappears in QA/PRD (Managed) after local pac code push to DV followed by PowerPlatform Pipeline redeploy

[HeritierLucMinani]

Describe the bug

Our Code App stops working in downstream environments (QA, PRD) after running pac code push locally against DV and then redeploying from DV (unmanaged) via the Power Platform Deployment Pipeline. The first deployment via PowerPlatform Pipeline works correctly. The issue only occurs on subsequent redeploys after running pac code push locallly.

Steps to Reproduce

1. Create a Code App in DV inside a solution.
2. Deploy the solution via PowerPlatform Pipeline (DV → QA/PRD) - app works correctly in QA/PRD.
3. Make code changes and run pac code push locally against DV - app works correctly in DV.
4. Trigger the PowerPlatform Pipeline again (DV → QA/PRD).
5. App is disappears in the environment we just deployed QA or PRD.

Expected behavior

The app should work in QA/PRD after each PP Pipeline redeploy, the same as it did on the first deployment.

Actual behavior

After step 4, the app disappears in QA/PRD. Running pac code push directly against QA/PRD to fix it is blocked because those environments contain managed solutions.

Screenshots or Error Messages

Pipeline used for deployment :

Application disappears from the environment

Environment information

• Framework, build tool or relevant package used: React, Vite, TypeScript
• Any connection/components: Dataverse (custom connector), Office 365 Users, SharePoint

Additional context

• 3 environments: DV, QA, PRD
• PP Pipeline configured as DV → QA(Managed) → PRD(Managed)
• Solution is deployed as managed to QA and PRD
• pac code push is used locally against DV only
• The Code App is included in the solution promoted via PP Pipeline
• When we redeploy one more time using the pipeline after the application disappeared, a new application is created but with a different UUID

[eschavez]

Thanks for the detailed report and the clear reproduction steps — this is really helpful!

We're investigating this. The behavior you're describing — where the app disappears in downstream managed environments after a pac code push + pipeline redeploy, and then reappears with a different UUID — suggests that pac code push may be altering the app's identity in the DV solution in a way that breaks the managed solution binding downstream.

A few things that would help us narrow this down faster:

1. After running pac code push against DV (step 3), can you export the solution from DV and check if the Code App component's GUID changed compared to the version before the push?
2. Are you exporting as unmanaged from DV and importing as managed downstream, or using the pipeline's native solution upgrade mechanism?
3. Does this also happen if you skip pac code push and instead make changes via the browser-based editor and push from there?

We'll update this issue as we learn more. Thanks again for the report!

[mickmal]

Hi @eschavez, I'm another dev working on this case.

Thanks for looking into this! Here is the information from our side:

1. Identity & GUID Check: I verified the solution files before and after running pac code push. The component identifier (cr09d_codeapppoc_c9adb) remains exactly the same—it is not generating a new ID upon push. Furthermore, the App ID found in the App Details also remains static. We have been able to use the exact same web link to access the app since its initial creation.
   (App Details reference)

2. Deployment Method: We are using the Power Platform Pipeline's native solution upgrade mechanism. We are not performing any manual export/import of unmanaged-to-managed solutions.

3. Browser-based Editor: Since this is a native Code App, the development workflow is strictly local. We didn't found a way to open or modify this specific component type via the standard Power Apps Studio browser editor (like you would with a traditional Canvas App), so we are unable to bypass the CLI to test that alternative route.

Let me know if there are any specific logs I can pull or other tests I can run to help narrow this down!

[mickmal]

As an additional update, on last Tuesday (2026.04.28), we also encountered a specific deployment error when trying to push the app to the QA environment after it had vanished for the first time.

The pipeline threw the following error:
ImportAsHolding failed with exception :Cannot add a Root Component cr09d_codeapppoc_c9adb of type 300 because it is not in the target system.

To fix the state and successfully deploy, we had to use the following workaround:

1. Delete the Code App from the Dev environment solution.
2. Deploy the solution to QA.
3. Add the Code App back into the Dev solution.
4. Deploy to QA again.

Sharing this in case the workaround steps help pinpoint how the component metadata is getting out of sync between the environments!