fix: improve max active requests calculation logic
- Use app model's max_active_requests or fall back to global config - When global APP_MAX_ACTIVE_REQUESTS is non-zero, enforce it as a hard limit - Add comprehensive docstring explaining the behavior Signed-off-by: -LAN- <laipz8200@outlook.com>pull/22847/head
parent
882f8bdd2c
commit
4c1c2cb90c
Loading…
Reference in New Issue