[metrics] Add in queue metrics (#4444)

This commit is contained in:
Stefan He
2025-04-09 17:19:27 -07:00
committed by GitHub
parent 4cb53ecd0c
commit 5db37c8626
3 changed files with 30 additions and 1 deletions

View File

@@ -474,6 +474,10 @@ class Req:
self.temp_scaled_logprobs = False
self.top_p_normalized_logprobs = False
# Latency Breakdown
self.queue_time_start = None
self.queue_time_end = None
# Logprobs (return values)
self.input_token_logprobs_val: Optional[List[float]] = None
self.input_token_logprobs_idx: Optional[List[int]] = None