Issue Description

llama-server /models api returns a 503 error while the model is loading. ramalama run handles this but ramalama chat currently does not. For interactive use this isn't a major issue but it would be good to fix this.

Worked around this for CI in #2342

Steps to reproduce the issue

Run ramalama chat soon after ramalama serve --detach

Describe the results you received

ramalama chat fails

Describe the results you expected

ramalama chat succeeds

ramalama info output

N/A

Upstream Latest Release

Yes

Additional environment details

No response

Additional information

No response