Retry when we get a generic LLM API error.

core/agents/base.py

@@ -135,7 +135,15 @@ class BaseAgent:
             )
             if answer.button == "yes":
                 return True
-
+        elif error == LLMError.GENERIC_API_ERROR:
+            await self.stream_handler(message)
+            answer = await self.ask_question(
+                "Would you like to retry the last step?",
+                buttons={"yes": "Yes", "no": "No"},
+                buttons_only=True,
+            )
+            if answer.button == "yes":
+                return True
         elif error == LLMError.RATE_LIMITED:
             await self.stream_handler(message)
 

core/llm/base.py

@@ -18,6 +18,7 @@ log = get_logger(__name__)
 class LLMError(str, Enum):
     KEY_EXPIRED = "key_expired"
     RATE_LIMITED = "rate_limited"
+    GENERIC_API_ERROR = "generic_api_error"
 
 
 class APIError(Exception):
@@ -240,6 +241,17 @@ class BaseLLMClient:
                 request_log.error = str(f"API error: {err}")
                 request_log.status = LLMRequestStatus.ERROR
                 return None, request_log
+            except (openai.APIError, anthropic.APIError, groq.APIError) as err:
+                # Generic LLM API error
+                # Make sure this handler is last in the chain as some of the above
+                # errors inherit from these `APIError` classes
+                log.warning(f"LLM API error {err}", exc_info=True)
+                msg = "LLM had an error processing our request."
+                if self.error_handler:
+                    should_retry = await self.error_handler(LLMError.GENERIC_API_ERROR, message=msg)
+                    if should_retry:
+                        continue
+                raise APIError(msg) from err
 
             request_log.response = response