Azure Functions Flex Consumption Python 3.12: successful deployment but zero functions register

Mark Green 0 Reputation points
2026-08-12T09:19:25.67+00:00

We have a Python 3.12 Azure Functions Flex Consumption application in UK South in a non-production development environment. Azure One Deploy with remote build completes successfully, but the Function host subsequently reports zero functions found/loaded and maps no HTTP routes.

The failure reproduces with two materially different packages:

  1. The reviewed full application, containing seven Python v2 HTTP functions. Local Python 3.12 discovery finds all seven functions.
  2. A deterministic three-file canary containing only function_app.py, host.json and requirements.txt, with one anonymous HTTP function and azure-functions as its only dependency. Local Python 3.12 discovery finds the function and direct invocation returns HTTP 200.

For both packages, Azure reports the remote build/deployment as complete, trigger synchronisation is requested, but the platform registers zero functions and every route returns 404. We also reapplied the existing Python 3.12 runtime and repeated the checks, with the same result.

We inspected the Azure-built full package and confirmed that the required files are at archive root, dependencies are present, and every reviewed source file matches byte-for-byte. Application Insights repeatedly records zero functions found/loaded, no job functions found and no HTTP routes mapped, with no application exception.

The Function App is stopped and there is no production impact or data loss. The issue blocks development integration and TEST/UAT.

This appears to require platform-side inspection of package activation/mount, Python worker startup/indexing, host storage access and trigger synchronisation. Please escalate this to a private Azure support ticket so we can securely provide the subscription, resource, deployment identifiers and sanitised diagnostic evidence bundle. We will not post customer-specific identifiers or diagnostic attachments publicly.

Azure Functions
Azure Functions

An Azure service that provides an event-driven serverless compute platform.


2 answers

Sort by: Most helpful
  1. SHOUMIK CHAKRAVARTY 0 Reputation points
    2026-08-14T04:13:17.96+00:00

    Hi @Mark Green ,

    A very similar case on this tag last week — 5969167— ended up being app‑side: the entry point was named main.py instead of function_app.py. Not your issue, since your canary already uses the right filename. He also linked #1808 as the same symptom with the correct filename, but that one’s closed with “needs author feedback,” from November 2025 on Python 3.11 — nothing much useful there.

    What hasn't been varied is the environment. Both suggestions so far changed the deployment mechanism, and you've done that twice. Three cheap tests:

    • Try another region — you've only tested UK South
    • Try Python 3.11 instead of 3.12
    • Spin up a brand‑new Function App with its own storage account

    If the canary works elsewhere, that points to a regional issue and gives you a much sharper support ticket. If it fails everywhere, that’s stronger evidence than a single environment.

    For the support ticket: Azure portal → Help + support -> Create a support request

    Was this answer helpful?

    0 comments No comments

  2. Allan Solomon Mejia 3,510 Reputation points
    2026-08-12T20:20:04.1266667+00:00

    Hello @Mark Green

    Since you've reproduced the same behavior with a minimal canary containing only function_app.py, host.json, requirements.txt, and azure-functions, I agree this is unlikely to be caused by application-level bindings or dependencies.

    One important point for Flex Consumption is that One Deploy is the only supported deployment technology. For Python, Microsoft also recommends packaging from the project root and using remote build.

    I would check Diagnose and solve problems → Flex Consumption Deployment in the Function App. Microsoft specifically provides this diagnostic for Flex Consumption and it can show deployment history, package status, and deployment-storage problems.

    I would also verify the configured deployment storage and its authentication:

    az functionapp config appsettings list \
      --name <function-app> \
      --resource-group <resource-group>
    

    and confirm the Function App can access the deployment storage container. Flex Consumption executes the deployed application from the package stored in that deployment container, so a successful build/upload doesn't necessarily prove that the host subsequently mounted and indexed the expected package.

    For one controlled test, I'd also deploy the minimal canary using the documented Core Tools path:

    func azure functionapp publish <APP_NAME>
    

    or the documented CLI remote-build path:

    az functionapp deployment source config-zip \
      --src <PACKAGE.zip> \
      --name <APP_NAME> \
      --resource-group <RESOURCE_GROUP> \
      --build-remote true
    

    Microsoft documents both for Python Flex Consumption deployments.

    If that still results in deployment succeeded → trigger sync completed → zero functions indexed, especially with the minimal canary, I think you've isolated this sufficiently for Azure Functions engineering review rather than further application troubleshooting.

    At that point, the deployment ID, UTC timestamp, Function host logs and sanitized output from the Flex Consumption Deployment diagnostic would be the most useful evidence for escalation.

    Please "Accept the Answer" if this information helped you. This will help us and others in the community as well.

    Was this answer helpful?


Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.