[Rust Frontend][Metrics] Add per-request preemption histogram (#57033)

Co-authored-by: Bugen Zhao <i@bugenzhao.com>
Signed-off-by: reidliu41 <reid201711@gmail.com>
This commit is contained in:
Reid
2026-09-16 02:51:17 +00:00
committed by GitHub
co-authored by Bugen Zhao
parent 7d4bba5498
commit 0673de9800
3 changed files with 33 additions and 0 deletions
+12
View File
@@ -790,6 +790,12 @@ async fn generate_records_request_metrics_in_prometheus_output() {
assert!(rendered.contains(&format!(
"vllm:num_preemptions_total{{model_name=\"{model_name}\",engine=\"4\"}} 1"
)));
assert!(rendered.contains(&format!(
"vllm:request_num_preemptions_sum{{model_name=\"{model_name}\",engine=\"4\"}} 1.0"
)));
assert!(rendered.contains(&format!(
"vllm:request_num_preemptions_count{{model_name=\"{model_name}\",engine=\"4\"}} 1"
)));
assert!(rendered.contains(&format!(
"vllm:time_to_first_token_seconds_count{{model_name=\"{model_name}\",engine=\"4\"}} 1"
)));
@@ -883,6 +889,12 @@ async fn dropping_stream_records_abort_terminal_request_metrics() {
assert!(rendered.contains(&format!(
"vllm:e2e_request_latency_seconds_count{{model_name=\"{model_name}\",engine=\"5\"}} 1"
)));
assert!(rendered.contains(&format!(
"vllm:request_num_preemptions_sum{{model_name=\"{model_name}\",engine=\"5\"}} 0.0"
)));
assert!(rendered.contains(&format!(
"vllm:request_num_preemptions_count{{model_name=\"{model_name}\",engine=\"5\"}} 1"
)));
llm.shutdown().await.unwrap();
}