feat(frontend): session id plumbing into requests (#48048)

Signed-off-by: Karen Chung <karenc@nvidia.com>
This commit is contained in:
Karen Chung
2026-08-03 22:18:58 +00:00
committed by GitHub
parent e578de311c
commit f57123aa2d
39 changed files with 438 additions and 4 deletions
@@ -215,6 +215,7 @@ class ServingTokens(GenerateBaseServing):
# Extract data_parallel_rank from header (router can inject it)
data_parallel_rank = self._get_data_parallel_rank(raw_request)
session_id = self._get_session_id_from_headers(raw_request)
result_generator = self.engine_client.generate(
engine_input,
@@ -224,6 +225,7 @@ class ServingTokens(GenerateBaseServing):
trace_headers=trace_headers,
priority=request.priority,
data_parallel_rank=data_parallel_rank,
session_id=session_id,
)
assert result_generator is not None