Derek Higgins

/lgtm

@stbenjam I see you approved the PR this was cherry picked from, can you do the same here We're trying figure out why the 4.10 baremetal job is failing while...

Hi @nateprewitt @sethmlarson , is this something you'd consider merging ?

@nateprewitt @sigmavirus24 @sethmlarson would you have a chance to look at this please, its been hanging around for over 2 years and still causing me problems.

Hi @frenzymadness , if you can review or confirm that this works for your use case then that would be appreciated but ultimately this PR has been waiting on reviews...

Proposed a fix here https://github.com/psf/requests/pull/5953

The experimental jobs failed, as expected as there was no cache or tokens configured https://github.com/meta-llama/llama-stack/actions/runs/15042121708/job/42276084080?pr=2176 See https://github.com/derekhiggins/llama-stack/actions/runs/15042046690/job/42275834971 For an example of this passing, with the cache, without hitting the provider...

Will fail because of lack of credentials (we would need to add these if merging) Example of new version passing here https://github.com/derekhiggins/llama-stack/actions/runs/15437347769/job/43446706608

Closing that the caching functionality being introduced here has been implemented by https://github.com/meta-llama/llama-stack/pull/2941

Just adding a bit more detail after some investigations The routes are called by [create_dynamic_typed_route](https://github.com/meta-llama/llama-stack/blob/7c0448456ed1dbca785606c8bde8797cb1c82704/llama_stack/distribution/server/server.py#L180) in llama_stack/distribution/server/server.py which calls them asynchronously with 'await' if the function is defined as a...