Unverified Commit e1f0835a authored by Mark McLoughlin's avatar Mark McLoughlin Committed by GitHub
Browse files

[V1][Metrics] Fix traceback with preemptions+LoRA (#14220)


Signed-off-by: default avatarMark McLoughlin <markmc@redhat.com>
parent 8ed5421a
......@@ -150,6 +150,7 @@ class IterationStats:
LoRARequestStates.scheduled_request(lora_stats, req_id)
elif event.type == EngineCoreEventType.PREEMPTED:
self.num_preempted_reqs += 1
LoRARequestStates.preempted_request(lora_stats, req_id)
def update_from_finished_request(self, finish_reason: "FinishReason",
num_prompt_tokens: int,
......@@ -224,6 +225,13 @@ class LoRARequestStates:
lora_stats.waiting_requests.remove(request_id)
lora_stats.running_requests.add(request_id)
@staticmethod
def preempted_request(lora_stats: Optional[LoRAStats], request_id: str):
if lora_stats is None:
return
lora_stats.running_requests.remove(request_id)
lora_stats.waiting_requests.add(request_id)
def update_iteration_stats(self,
iteration_stats: Optional[IterationStats]):
if iteration_stats is None:
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment