From 9545b0e5cde2244c1e86416c5c57dfbbcbca3737 Mon Sep 17 00:00:00 2001
From: Ishaan Jaff <ishaanjaffer0324@gmail.com>
Date: Fri, 1 Nov 2024 18:36:17 +0530
Subject: [PATCH] (fix) slack alerting - don't spam the failed cost tracking
 alert for the same model  (#6543)

* fix use failing_model as cache key for failed_tracking_alert

* fix use standard logging payload for getting response cost

* fix  kwargs.get("response_cost")

* fix getting response cost
---
 litellm/integrations/SlackAlerting/slack_alerting.py | 12 +++++++++---
 litellm/proxy/proxy_config.yaml                      | 10 +++++-----
 litellm/proxy/proxy_server.py                        |  5 +++--
 litellm/proxy/utils.py                               |  9 +++++++--
 4 files changed, 24 insertions(+), 12 deletions(-)

diff --git a/litellm/integrations/SlackAlerting/slack_alerting.py b/litellm/integrations/SlackAlerting/slack_alerting.py
index dbe9e4161..85d54a337 100644
--- a/litellm/integrations/SlackAlerting/slack_alerting.py
+++ b/litellm/integrations/SlackAlerting/slack_alerting.py
@@ -550,8 +550,14 @@ class SlackAlerting(CustomBatchLogger):
                     alerting_metadata=alerting_metadata,
                 )
 
-    async def failed_tracking_alert(self, error_message: str):
-        """Raise alert when tracking failed for specific model"""
+    async def failed_tracking_alert(self, error_message: str, failing_model: str):
+        """
+        Raise alert when tracking failed for specific model
+
+        Args:
+            error_message (str): Error message
+            failing_model (str): Model that failed tracking
+        """
         if self.alerting is None or self.alert_types is None:
             # do nothing if alerting is not switched on
             return
@@ -560,7 +566,7 @@ class SlackAlerting(CustomBatchLogger):
 
         _cache: DualCache = self.internal_usage_cache
         message = "Failed Tracking Cost for " + error_message
-        _cache_key = "budget_alerts:failed_tracking:{}".format(message)
+        _cache_key = "budget_alerts:failed_tracking:{}".format(failing_model)
         result = await _cache.async_get_cache(key=_cache_key)
         if result is None:
             await self.send_alert(
diff --git a/litellm/proxy/proxy_config.yaml b/litellm/proxy/proxy_config.yaml
index 95eff095c..23834f759 100644
--- a/litellm/proxy/proxy_config.yaml
+++ b/litellm/proxy/proxy_config.yaml
@@ -1,11 +1,11 @@
 model_list:
   - model_name: gpt-4o
     litellm_params:
-      model: gpt-4o
+      model: openai/gpt-5
       api_key: os.environ/OPENAI_API_KEY
       api_base: https://exampleopenaiendpoint-production.up.railway.app/
 
-litellm_settings:
-  callbacks: ["prometheus"]
-  service_callback: ["prometheus_system"]
-  cache: true
+
+general_settings: 
+    alerting: ["slack"]
+    alerting_threshold: 0.001
diff --git a/litellm/proxy/proxy_server.py b/litellm/proxy/proxy_server.py
index 1f0271f34..ca6befef6 100644
--- a/litellm/proxy/proxy_server.py
+++ b/litellm/proxy/proxy_server.py
@@ -265,6 +265,7 @@ from litellm.types.llms.anthropic import (
 )
 from litellm.types.llms.openai import HttpxBinaryResponseContent
 from litellm.types.router import RouterGeneralSettings
+from litellm.types.utils import StandardLoggingPayload
 
 try:
     from litellm._version import version
@@ -778,7 +779,6 @@ async def _PROXY_track_cost_callback(
         if kwargs.get("response_cost", None) is not None:
             response_cost = kwargs["response_cost"]
             user_api_key = metadata.get("user_api_key", None)
-
             if kwargs.get("cache_hit", False) is True:
                 response_cost = 0.0
                 verbose_proxy_logger.info(
@@ -838,13 +838,14 @@ async def _PROXY_track_cost_callback(
                     f"Cost tracking failed for model={model}.\nDebug info - {cost_tracking_failure_debug_info}\nAdd custom pricing - https://docs.litellm.ai/docs/proxy/custom_pricing"
                 )
     except Exception as e:
-        error_msg = f"error in tracking cost callback - {traceback.format_exc()}"
+        error_msg = f"Error in tracking cost callback - {str(e)}\n Traceback:{traceback.format_exc()}"
         model = kwargs.get("model", "")
         metadata = kwargs.get("litellm_params", {}).get("metadata", {})
         error_msg += f"\n Args to _PROXY_track_cost_callback\n model: {model}\n metadata: {metadata}\n"
         asyncio.create_task(
             proxy_logging_obj.failed_tracking_alert(
                 error_message=error_msg,
+                failing_model=model,
             )
         )
         verbose_proxy_logger.debug("error in tracking cost callback - %s", e)
diff --git a/litellm/proxy/utils.py b/litellm/proxy/utils.py
index 656e97a6c..8919da978 100644
--- a/litellm/proxy/utils.py
+++ b/litellm/proxy/utils.py
@@ -667,13 +667,18 @@ class ProxyLogging:
                 raise e
         return data
 
-    async def failed_tracking_alert(self, error_message: str):
+    async def failed_tracking_alert(
+        self,
+        error_message: str,
+        failing_model: str,
+    ):
         if self.alerting is None:
             return
 
         if self.slack_alerting_instance:
             await self.slack_alerting_instance.failed_tracking_alert(
-                error_message=error_message
+                error_message=error_message,
+                failing_model=failing_model,
             )
 
     async def budget_alerts(