Merge branch 'main' into litellm_add_bedrock_guardrails

litellm/proxy/_new_secret_config.yaml

@@ -2,12 +2,3 @@ model_list:
   - model_name: "*"
     litellm_params:
       model: "*"
-
-litellm_settings:
-  success_callback: ["s3"]
-  cache: true
-  s3_callback_params:
-    s3_bucket_name: mytestbucketlitellm   # AWS Bucket Name for S3
-    s3_region_name: us-west-2              # AWS Region Name for S3
-    s3_aws_access_key_id: os.environ/AWS_ACCESS_KEY_ID  # us os.environ/<variable name> to pass environment variables. This is AWS Access Key ID for S3
-    s3_aws_secret_access_key: os.environ/AWS_SECRET_ACCESS_KEY  # AWS Secret Access Key for S3

litellm/proxy/auth/auth_checks.py

@@ -66,7 +66,7 @@ def common_checks(
         raise Exception(
             f"Team={team_object.team_id} is blocked. Update via `/team/unblock` if your admin."
         )
-    # 2. If user can call model
+    # 2. If team can call model
     if (
         _model is not None
         and team_object is not None
@@ -74,7 +74,11 @@ def common_checks(
         and _model not in team_object.models
     ):
         # this means the team has access to all models on the proxy
-        if "all-proxy-models" in team_object.models:
+        if (
+            "all-proxy-models" in team_object.models
+            or "*" in team_object.models
+            or "openai/*" in team_object.models
+        ):
             # this means the team has access to all models on the proxy
             pass
         # check if the team model is an access_group

litellm/proxy/proxy_config.yaml

@@ -1,8 +1,9 @@
 model_list:
   - model_name: fake-openai-endpoint
     litellm_params:
-      model: openai/gpt-4
-      api_key: os.environ/OPENAI_API_KEY
+      model: openai/fake
+      api_key: fake-key
+      api_base: https://exampleopenaiendpoint-production.up.railway.app/
 
 guardrails:
   - guardrail_name: "bedrock-pre-guard"

litellm/proxy/proxy_server.py

@@ -1588,7 +1588,7 @@ class ProxyConfig:
                         verbose_proxy_logger.debug(  # noqa
                             f"{blue_color_code}Set Cache on LiteLLM Proxy: {vars(litellm.cache.cache)}{reset_color_code}"
                         )
-                elif key == "cache" and value == False:
+                elif key == "cache" and value is False:
                     pass
                 elif key == "guardrails":
                     if premium_user is not True:
@@ -2672,6 +2672,13 @@ def giveup(e):
         and isinstance(e.message, str)
         and "Max parallel request limit reached" in e.message
     )
+
+    if (
+        general_settings.get("disable_retry_on_max_parallel_request_limit_error")
+        is True
+    ):
+        return True  # giveup if queuing max parallel request limits is disabled
+
     if result:
         verbose_proxy_logger.info(json.dumps({"event": "giveup", "exception": str(e)}))
     return result