imbajin commented on PR #358:
URL: https://github.com/apache/hugegraph-ai/pull/358#issuecomment-4610137209

   Follow-up note on `request_graph_config()` concurrency:
   
   The current snapshot/restore is only a transitional guard. It reduces 
permanent global config leakage, but it does not provide true concurrent 
request isolation because `/rag`, `/rag/graph`, and `/text2gremlin` still 
temporarily mutate process-global `huge_settings`.
   
   I do not think we should fix this by adding a global lock in this PR. That 
critical section covers potentially long-running RAG/Text2Gremlin execution, 
including graph queries, vector retrieval, and LLM calls; locking it would 
effectively serialize concurrent requests.
   
   Suggested scope for this PR:
   - keep the snapshot/restore improvement;
   - strengthen the comment to an explicit `FIXME`;
   - leave the real fix to a dedicated follow-up PR.
   
   Expected follow-up direction: pass graph config through request-scoped 
flow/operator context and create per-request HugeGraph clients instead of 
reading/writing `huge_settings` globally.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to