[Adapter for Container Model] - Depricated Gradle Features Causing a Tag Release Fail

Ask the Community
,
1

I’m working in the Models app & Code Repository (Adapter Template) on Palantir Foundry, attempting to configure a container-backed model adapter for my model. I have:

  • Pushed the image to Palantir’s docker hub (container model configured).
  • Created a code repository using the Model Adapter Library template.
  • Implemented my adapter logic and comitted to remote.
  • While I create a release tag, the build is getting failed because of an depricated gradle. For which I don’t have an control of.

Note: Have came to know that “foundry-ml-api“ was depticated from oct/2025. Resolved the “Could not find the matchspec in any channel“ error by removing the depedencies from HAWK file.

Error Log From Tag Release:

“““

Deprecated Gradle features were used in this build, making it incompatible with Gradle 9.0.

BUILD SUCCESSFUL in 1m 48s
27 actionable tasks: 27 executed
An error occurred while applying the patch: Default:Internal. Error instance id: e53ac830-4be0-4a7f-a6d4-c1335f10e2f7.

“““

Build Summary

“““

:condaPackRun                               | SUCCESS    | 37.729 secs
:runVersions                                | SUCCESS    | 18.644 secs
:setupConda                                 | SUCCESS    | 9.413 secs
:createCondaPackLayer                       | SUCCESS    | 7.357 secs
:condaBuild                                 | SUCCESS    | 5.153 secs
:uploadLayersToArtifactsDocker              | SUCCESS    | 4.066 secs
:createModelAdapterManifest                 | SUCCESS    | 3.714 secs
:distEgg                                    | SUCCESS    | 0.943 secs
:condaLibraryPublish                        | SUCCESS    | 0.573 secs
:registerModelAdapter                       | SUCCESS    | 0.55 secs
:setupMaestro                               | SUCCESS    | 0.471 secs
:condaRender                                | SUCCESS    | 0.305 secs
:setupHawk                                  | SUCCESS    | 0.201 secs
:checkBackingRepositoriesPermissions        | SUCCESS    | 0.19 secs
:condaRunList                               | SUCCESS    | 0.138 secs
:convertTrustStore                          | SUCCESS    | 0.085 secs
:createModelAdapterLayer                    | SUCCESS    | 0.061 secs
:symlinkPythonExecutable                    | SUCCESS    | 0.06 secs
:checkStaticInstaller                       | SUCCESS    | 0.032 secs
:versionPy                                  | SUCCESS    | 0.024 secs
:runRequires                                | SUCCESS    | 0.02 secs
:condaLocksPatch                            | SUCCESS    | 0.017 secs
:symlinkMambaExecutable                     | SUCCESS    | 0.016 secs
:setupPythonEnv                             | SKIPPED    | 0.011 secs
:symlinkCondaExecutable                     | SUCCESS    | 0.01 secs
:condaConfigureNewBackingRepositories       | SUCCESS    | 0.009 secs
:generateRepoPy                             | SUCCESS    | 0.004 secs
:condaConfigureNewBackingRepositoriesPatch  | SUCCESS    | 0.004 secs
:condaInfo                                  | SKIPPED    | 0.002 secs
:extractCondaWrapper                        | SKIPPED    | 0.001 secs
:checkSharedLibraries                       | SKIPPED    | 0.001 secs
:runPrintLockFile                           | SKIPPED    | 0.0 secs
:check                                      | UNKNOWN    | 0.0 secs
:prepackage                                 | UNKNOWN    | 0.0 secs
:patch                                      | UNKNOWN    | 0.0 secs
:publish                                    | UNKNOWN    | 0.0 secs

“““

Please share some knowledge if you guys ended-up with simear issues. Thanks in advance.

2

Hi there!

Just checking in to make sure you were able to resolve your issue. From your note it sounds like that may be the case!

3

Unfortunately no, the issue still persist. Crawling through documentation to find a solution.

4

Hi @kevindaniel ,

The deprecated gradle feature log line is probably a red herring and does not necessarily indicate a failure - it’s just a warning. Looking at the Build summary, no tasks seem to be in a failed state, while the release tag says “BUILD SUCCESSFUL in 1m 48s”. Have you tried going through the model creation flow and confirming whether or not an adapter exists from your repository based on the steps described here?

Thanks,

Julien