Refactor rate limit handling (#280)

* - Refactored retry for rate limits into the base class
- ConsumerProducer is derived from Consumer to simplify code
- Added rate_limit_count metrics for rate limit events

* Add rate limit events to VertexAI and Google AI Studio

* Added Grafana rate limit dashboard

* Add rate limit handling to all LLMs

trustgraph-vertexai/trustgraph/model/text_completion/vertexai/llm.py

@@ -178,25 +178,15 @@ class Processor(ConsumerProducer):
 
         except google.api_core.exceptions.ResourceExhausted as e:
 
-            print("Send rate limit response...", flush=True)
+            print("Hit rate limit:", e, flush=True)
 
-            r = TextCompletionResponse(
-                error=Error(
-                    type = "rate-limit",
-                    message = str(e),
-                ),
-                response=None,
-                in_token=None,
-                out_token=None,
-                model=None,
-            )
-
-            self.producer.send(r, properties={"id": id})
-
-            self.consumer.acknowledge(msg)
+            # Leave rate limit retries to the base handler
+            raise TooManyRequests()
 
         except Exception as e:
 
+            # Apart from rate limits, treat all exceptions as unrecoverable
+
             print(f"Exception: {e}")
 
             print("Send error response...", flush=True)