Misc fix for min_p_sampling, --cuda-graph-bs (#2761)

python/sglang/srt/metrics/collector.py

@@ -114,26 +114,20 @@ class TokenizerMetricsCollector:
             documentation="Histogram of time to first token in seconds.",
             labelnames=labels.keys(),
             buckets=[
-                0.001,
-                0.005,
-                0.01,
-                0.02,
-                0.04,
-                0.06,
-                0.08,
                 0.1,
                 0.25,
                 0.5,
                 0.75,
-                1.0,
-                2.5,
-                5.0,
-                7.5,
-                10.0,
-                15.0,
-                20.0,
-                25.0,
-                30.0,
+                1,
+                2,
+                5,
+                10,
+                20,
+                40,
+                60,
+                80,
+                120,
+                160,
             ],
         )
 
@@ -168,21 +162,19 @@ class TokenizerMetricsCollector:
             documentation="Histogram of End-to-end request latency in seconds",
             labelnames=labels.keys(),
             buckets=[
-                0.3,
+                0.1,
+                0.25,
                 0.5,
-                0.8,
-                1.0,
-                1.5,
-                2.0,
-                2.5,
-                5.0,
-                10.0,
-                15.0,
-                20.0,
-                30.0,
-                40.0,
-                50.0,
-                60.0,
+                1,
+                2,
+                5,
+                10,
+                20,
+                40,
+                60,
+                80,
+                120,
+                160,
             ],
         )