Skip to content

Commit 6092d8a

Browse files
committed
fix(metrics): restore kv cache hit-rate name
1 parent 2f02820 commit 6092d8a

2 files changed

Lines changed: 5 additions & 0 deletions

File tree

lib/bindings/python/src/dynamo/prometheus_names.py

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -271,6 +271,8 @@ class kvstats:
271271
TOTAL_BLOCKS = "total_blocks"
272272
# GPU cache usage as a percentage (0.0-1.0)
273273
GPU_CACHE_USAGE_PERCENT = "gpu_cache_usage_percent"
274+
# Prefix cache hit rate (0.0-1.0), portable across vLLM / SGLang / TRT-LLM
275+
KV_CACHE_HIT_RATE = "kv_cache_hit_rate"
274276

275277

276278
class labels:

lib/runtime/src/metrics/prometheus_names.rs

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -823,6 +823,9 @@ pub mod kvstats {
823823

824824
/// GPU cache usage as a percentage (0.0-1.0)
825825
pub const GPU_CACHE_USAGE_PERCENT: &str = "gpu_cache_usage_percent";
826+
827+
/// Prefix cache hit rate (0.0-1.0), portable across vLLM / SGLang / TRT-LLM
828+
pub const KV_CACHE_HIT_RATE: &str = "kv_cache_hit_rate";
826829
}
827830

828831
// Model information metrics

0 commit comments

Comments
 (0)