diff --git a/ui/litellm-dashboard/src/components/model_dashboard.tsx b/ui/litellm-dashboard/src/components/model_dashboard.tsx index c63f6ed93..9d1c6b6c2 100644 --- a/ui/litellm-dashboard/src/components/model_dashboard.tsx +++ b/ui/litellm-dashboard/src/components/model_dashboard.tsx @@ -457,8 +457,8 @@ const handleEditSubmit = async (formValues: Record) => { ); console.log("Model metrics response:", modelMetricsResponse); - // Sort by latency (avg_latency_seconds) - const sortedByLatency = [...modelMetricsResponse].sort((a, b) => b.avg_latency_seconds - a.avg_latency_seconds); + // Sort by latency (avg_latency_per_token) + const sortedByLatency = [...modelMetricsResponse].sort((a, b) => b.avg_latency_per_token - a.avg_latency_per_token); console.log("Sorted by latency:", sortedByLatency); setModelMetrics(modelMetricsResponse); @@ -686,8 +686,8 @@ const handleEditSubmit = async (formValues: Record) => {
All Models Add Model - Model Analytics
/health Models
+ Model Analytics
@@ -992,6 +992,17 @@ const handleEditSubmit = async (formValues: Record) => { + + + `/health` will run a very small request through your models configured on litellm + + + {healthCheckResponse && ( +
{JSON.stringify(healthCheckResponse, null, 2)}
+ )} + +
+

View how requests were load balanced within a model group

(Beta feature) only supported for Azure Model Groups

@@ -1017,20 +1028,33 @@ const handleEditSubmit = async (formValues: Record) => { ))} - - Number Requests per Model - - - + + + + + + + Model + Median Latency/Token + + + + {modelLatencyMetrics.map((metric, idx) => ( + + {metric.model} + {metric.avg_latency_per_token.toFixed(4)} + + ))} + +
+
+ + + + Requests, Failures per Model + + +
Exceptions per Model ) => { /> - - Latency Per Model - - +
- - - `/health` will run a very small request through your models configured on litellm - - - {healthCheckResponse && ( -
{JSON.stringify(healthCheckResponse, null, 2)}
- )} - -
-
+