fix(main.py): for health checks, don't use cached responses

2025-04-25 10:44:24 +00:00 · 2024-02-02 16:51:42 -08:00 · 2024-02-02 16:51:42 -08:00 · 62ad6f19b7
commit 62ad6f19b7
parent d69edac11b
2 changed files with 22 additions and 0 deletions
--- a/litellm/main.py
+++ b/litellm/main.py
@ -3259,6 +3259,9 @@ async def ahealth_check(
                organization=organization,
            )
        else:
+            model_params["cache"] = {
+                "no-cache": True
+            }  # don't used cached responses for making health check calls
            if mode == "embedding":
                model_params.pop("messages", None)
                model_params["input"] = input