Workaround for the MCP server connection cleanup issue

[TamiTakamiya]

Description

This is a workaround for the MCP server issue found by @ldjebran while we are working on #52. Refer to this PR comment for more details.

It is required for Llama Stack 0.4.3. We should be able to take this workaround out when we move to Llama Stack 0.5.x.

Type of change

• Refactor
• New feature
• Bug fix
• CVE fix
• Optimization
• Documentation Update
• Configuration Update
• Bump-up service version
• Bump-up dependent library
• Bump-up library or tool used for development (does not change the final image)
• CI configuration change
• Unit tests improvement

Tools used to create PR

Identify any AI code assistants used in this PR (for transparency and review context)

• Assisted-by: Claude Code
• Generated by: Claude Code

Related Tickets & Documents

• Related Issue # AAP-66984
• Closes # AAP-66984

Checklist before requesting a review

• I have performed a self-review of my code.
• PR has passed all pre-merge test jobs.
• If it is a core feature, I have added thorough tests.

Testing

• Please provide detailed steps to perform tests related to this code change.

Unit test cases are provided. For manual testing, configure a RAG chatbot with MCP enabled (any MCP server that uses a network connection should be fine) and issue a streaming quey API.

• How were the fix/results from this change verified? Please provide relevant screenshots or results.

Manual verification can be done using Ansible Lightspeed Intelligent Agent (ALIA) BYOK feature, which are being built using Lightspeed Core Stack 0.4.x + Llama Stack 0.4.3.

[ldjebran]

LGTM

[jrobertboos]

It is required for Llama Stack 0.4.3. We should be able to take this workaround out when we move to Llama Stack 0.5.x.

I am currently working on bumping both this repo and lightspeed-stack to 0.5.x which will hopefully be done by Wednesday at the latest. So if you want to wait that could be option... Totally understand if not though :)

[TamiTakamiya]

It is required for Llama Stack 0.4.3. We should be able to take this workaround out when we move to Llama Stack 0.5.x.

I am currently working on bumping both this repo and lightspeed-stack to 0.5.x which will hopefully be done by Wednesday at the latest. So if you want to wait that could be option... Totally understand if not though :)

@jrobertboos We (Ansible Lightspeed team) is planning to release our next version based on the coming lightspeed-stack 0.4.2. So we do need to have this workaround for our release. After the lightspeed-providers 0.4.2 is released with this workaround, please update this repo to Llama Stack 0.5.x and take out this workaround. Thanks.