iHore/vLLM
8
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

[Doc][Metrics] Fix spec-decode PromQL examples to use the exposed names (#55016)

Browse Source 
Signed-off-by: Michele Campi <michele.campi@outlook.com>
This commit is contained in:
Michele Campi
2026-09-14 21:15:28 +00:00
committed by GitHub
parent dabc4362b4
commit d392ac836e
1 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
vllm/v1/spec_decode/metrics.py
+3 -3
View File
@@ -187,12 +187,12 @@ class SpecDecodingProm:
1 + (
rate(vllm:spec_decode_num_accepted_tokens_total[$interval]) /
rate(vllm:spec_decode_num_drafts[$interval]))
rate(vllm:spec_decode_num_drafts_total[$interval]))
A per-position acceptance rate vector can be computed using
vllm:spec_decode_num_accepted_tokens_per_pos[$interval] /
vllm:spec_decode_num_drafts[$interval]
vllm:spec_decode_num_accepted_tokens_per_pos_total[$interval] /
vllm:spec_decode_num_drafts_total[$interval]
"""
_counter_cls = prometheus_client.Counter