Description
QA needs to test LLM request priority in a self-hosted environment. The feature spans the LLM stack (llm-api-gateway, llm-request-router), the NVCF API, and a Cassandra schema migration, but the self-managed stack still pins older versions, so a stack install today cannot exercise it.
Wire the current LLM stack, NVCF API, and Cassandra migration into the self-managed stack 0.7 release so request priority can be tested end to end on a self-hosted install.
Definition of Done
- The self-managed stack deploys the LLM stack, NVCF API, and Cassandra migrations at versions that carry the request-priority changes.
- The version catalog and published manifest match what the stack deploys.
- A fresh install and an upgrade from the previous stack release both come up cleanly with the LLM addon enabled.
Resources
- Stack pins:
deploy/stacks/self-managed/helmfile.d/
- Version catalog:
docs/version-catalog/main.yaml
Description
QA needs to test LLM request priority in a self-hosted environment. The feature spans the LLM stack (llm-api-gateway, llm-request-router), the NVCF API, and a Cassandra schema migration, but the self-managed stack still pins older versions, so a stack install today cannot exercise it.
Wire the current LLM stack, NVCF API, and Cassandra migration into the self-managed stack 0.7 release so request priority can be tested end to end on a self-hosted install.
Definition of Done
Resources
deploy/stacks/self-managed/helmfile.d/docs/version-catalog/main.yaml