How can I verify the currently mounted Azure Functions Flex Consumption package?

Tim McConnehey 0 Reputation points
2026-11-10T22:55:13.0566667+00:00

For an Azure Functions Flex Consumption app using package deployment, what supported read-only method can confirm the package currently mounted by the running app when Deployment Center has no release rows and the Flex Consumption Deployment diagnostic only shows recent activity? Can the exact active deployment ID be associated with an immutable package digest or storage version using existing read access, with public and deployment-management ingress left closed?

Azure Functions
Azure Functions

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


1 answer

Sort by: Most helpful
  1. Vinodh247-1375 44,281 Reputation points Volunteer Moderator
    2026-11-15T16:37:36.37+00:00

    The important distinction here is between identifying the deployment artefact in storage and obtaining authoritative runtime attestation of the package being executed.

    For a Flex Consumption function app, the deployment storage configuration can be read from properties.functionAppConfig.deployment.storage. Azure Blob Storage is the deployment source for apps running on the Flex Consumption plan. [learn.microsoft.com], [learn.microsoft.com]

    With read access to that storage account, you can establish:

    • which deployment container is configured;
    • the blob name, ETag, last-modified time and version ID, when versioning is enabled;
    • the SHA-256 digest of the ZIP downloaded from that specific blob version.

    This provides a verifiable identity for the stored artefact, but it does not authoritatively prove that a running worker has mounted that exact blob version.

    The current documented Flex Consumption management surface does not expose a read-only runtime property that correlates all of the following:

    deployment ID     -> blob version ID     -> package digest     -> package mounted by the running worker

    Deployment Center and the Flex Consumption deployment diagnostic describe deployment activity and status, but they are not a runtime package-attestation interface. The existing thread response reaches the same conclusion, so inspecting another ARM property or recalculating the current blob hash would not close this correlation gap. [learn.microsoft.com]

    Also, a Blob version ID should not automatically be described as immutable. Blob versioning preserves distinct versions, but protection against modification or deletion requires an Azure Blob Storage immutability policy or an equivalent retention control.

    Therefore, with the available read permissions and with public and deployment-management ingress closed, the supported conclusion is:

    You can identify and hash the package currently present in the configured deployment storage, but you cannot conclusively attest through a documented read-only API that a particular running Flex Consumption worker has mounted that exact package version.

    For future deployments, create the correlation before publishing:

    1. Calculate the ZIP SHA-256 in the deployment pipeline.
    2. Record the Git commit, pipeline run ID, package digest and resulting Blob version ID in an append-only deployment manifest.
    3. Include the same release identifier inside the package.
    4. Emit it through Application Insights as a startup custom event or trace.

    This does not make the platform expose a mounted-package identifier, but it creates a practical evidence chain between the build, stored artefact and observed application startup without opening SCM/Kudu or public deployment ingress. Azure Functions supports production deployment through external pipelines such as Azure Pipelines and GitHub Actions, which are suitable places to generate this evidence. [learn.microsoft.com]

    References:

    Help make this community better for everyone: if this answer resolved your issue, please accept it or leave an upvote. If not, share more details in a comment so we can continue the discussion and find the right solution

    Was this answer helpful?

    0 comments No comments

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.