fix t_llm_request is unbounded problem (#657)

skyvern/forge/sdk/api/llm/api_handler_factory.py

@@ -190,12 +190,12 @@ class LLMAPIHandlerFactory:
                         }
                     ).encode("utf-8"),
                 )
+            t_llm_request = time.perf_counter()
             try:
                 # TODO (kerem): add a timeout to this call
                 # TODO (kerem): add a retry mechanism to this call (acompletion_with_retries)
                 # TODO (kerem): use litellm fallbacks? https://litellm.vercel.app/docs/tutorials/fallbacks#how-does-completion_with_fallbacks-work
                 LOG.info("Calling LLM API", llm_key=llm_key, model=llm_config.model_name)
-                t_llm_request = time.perf_counter()
                 response = await litellm.acompletion(
                     model=llm_config.model_name,
                     messages=messages,