Skip to content

fix: async llm engine didnt have get_metrics() #1943

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
terrykong merged 2 commits into from
Feb 14, 2026
Merged

fix: async llm engine didnt have get_metrics() #1943

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
9945ffc
fix: async llm engine didnt have get_metrics()
terrykong Feb 13, 2026
be2df01
super silly mistake
terrykong Feb 13, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 8 additions & 1 deletion nemo_rl/models/generation/vllm/vllm_worker.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -539,7 +539,14 @@ def _get_raw_spec_counters(self) -> dict[str, float | list[float]]:
"""
metrics: dict[str, float | list[float]] = {}
if self.llm is not None:
for metric in self.llm.get_metrics():
if hasattr(self.llm, "get_metrics"):
vllm_prom_metrics = self.llm.get_metrics()
else:
# The AsyncLLM API does not implement get_metrics so we need to call the prometheus API ourselves
from vllm.v1.metrics.reader import get_metrics_snapshot

vllm_prom_metrics = get_metrics_snapshot()
for metric in vllm_prom_metrics:
if hasattr(metric, "values"):
metrics[metric.name] = metric.values
elif hasattr(metric, "value"):
Expand Down
1 change: 1 addition & 0 deletions tests/functional/grpo_non_colocated.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,6 +27,7 @@ uv run coverage run -a --data-file=$PROJECT_ROOT/tests/.coverage --source=$PROJE
policy.train_micro_batch_size=1 \
policy.generation.colocated.enabled=false \
policy.generation.colocated.resources.gpus_per_node=1 \
policy.generation.vllm_cfg.async_engine=true \
cluster.gpus_per_node=2 \
grpo.max_num_steps=2 \
logger.tensorboard_enabled=true \
Expand Down
Loading