phoenix/litellm-mirror
0
0
Fork 1
mirror of https://github.com/BerriAI/litellm.git synced 2025-04-27 11:43:54 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

prometheus log_success_fallback_event

Browse source
This commit is contained in:
Ishaan Jaff 2024-08-10 14:05:18 -07:00
parent ada3ae670b
commit 3ecf4db741
2 changed files with 34 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

33
litellm/integrations/prometheus.py
View file

@ -170,6 +170,17 @@ class PrometheusLogger(CustomLogger):
labelnames=_logged_llm_labels, labelnames=_logged_llm_labels,
) )
self.llm_deployment_successful_fallbacks = Counter(
"llm_deployment_successful_fallbacks",
"LLM Deployment Analytics - Number of successful fallback workloads",
["primary_model", "fallback_model"],
)
self.llm_deployment_failed_fallbacks = Counter(
"llm_deployment_failed_fallbacks",
"LLM Deployment Analytics - Number of failed fallback workloads",
["primary_model", "fallback_model"],
)
except Exception as e: except Exception as e:
print_verbose(f"Got exception on init prometheus client {str(e)}") print_verbose(f"Got exception on init prometheus client {str(e)}")
raise e raise e
@ -479,6 +490,28 @@ class PrometheusLogger(CustomLogger):
) )
return return
async def log_success_fallback_event(self, original_model_group: str, kwargs: dict):
verbose_logger.debug(
"Prometheus: log_success_fallback_event, original_model_group: %s, kwargs: %s",
original_model_group,
kwargs,
)
_new_model = kwargs.get("model")
self.llm_deployment_successful_fallbacks.labels(
primary_model=original_model_group, fallback_model=_new_model
).inc()
async def log_failure_fallback_event(self, original_model_group: str, kwargs: dict):
verbose_logger.debug(
"Prometheus: log_failure_fallback_event, original_model_group: %s, kwargs: %s",
original_model_group,
kwargs,
)
_new_model = kwargs.get("model")
self.llm_deployment_failed_fallbacks.labels(
primary_model=original_model_group, fallback_model=_new_model
).inc()
def set_deployment_state( def set_deployment_state(
self, self,
state: int, state: int,

1
litellm/proxy/proxy_config.yaml
View file

@ -37,5 +37,6 @@ general_settings:
master_key: sk-1234 master_key: sk-1234
litellm_settings: litellm_settings:
fallbacks: [{"gemini-1.5-pro-001": ["gpt-4o"]}]
success_callback: ["langfuse", "prometheus"] success_callback: ["langfuse", "prometheus"]
failure_callback: ["prometheus"] failure_callback: ["prometheus"]