[CI] Pass the scale-out endpoint flag in EC E2E (#56819)
Signed-off-by: Kevin Luu <51931015+khluu@users.noreply.github.com> Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com> Co-authored-by: OpenAI Codex <codex@openai.com> Co-authored-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
Kevin H. Luu committed
Sep 14, 2026 at 21:25 UTC
caf56ce9dac8614b11d682d1cfd288cfbe473a43
2 files changed
+5
-3
docs/usage/security.md
+1
-1
@@ -170,7 +170,7 @@ When `--api-key` is configured, the following endpoints require Bearer token aut
170
- `/v1/messages/render` - Render Anthropic-compatible messages (available on `vllm serve` only when `--enable-scale-out` is set, or on `vllm launch render`)
171
- `/v1/messages/count_tokens` - Count tokens for Anthropic messages
172
- `/v1/responses` - Create a response
173
-- `/v1/responses/render` - Render a self-contained response request (available on `vllm serve` only when `VLLM_ENABLE_SCALE_OUT_ENDPOINTS=1`, or on `vllm launch render` unless explicitly disabled)
173
+- `/v1/responses/render` - Render a self-contained response request (available on `vllm serve` only when `--enable-scale-out` is set, or on `vllm launch render` unless explicitly disabled)
174
- `/v1/responses/{response_id}` - Retrieve a response
175
- `/v1/responses/{response_id}/cancel` - Cancel a response
176
- `/v1/score` - Scoring API
tests/entrypoints/scale_out/ec_integration/run_scale_out_ec_e2e_test.sh
+4
-2
@@ -181,8 +181,9 @@ run_scale_out_ec() {
181
# Encode-only EC producer instance. It runs the vision encoder and
182
# publishes embeddings through the ECExampleConnector shared storage.
183
echo "Starting encode instance on GPU $GPU_E, port $ENCODE_PORT"
184
- env CUDA_VISIBLE_DEVICES="$GPU_E" VLLM_ENABLE_SCALE_OUT_ENDPOINTS=1 \
184
+ env CUDA_VISIBLE_DEVICES="$GPU_E" \
185
vllm serve "$MODEL" \
186
+ --enable-scale-out \
187
--port "$ENCODE_PORT" \
188
--max-model-len "$MAX_MODEL_LEN" \
189
--enforce-eager \
@@ -204,8 +205,9 @@ run_scale_out_ec() {
205
# features plus ec_transfer_params and loads the embeddings that the
206
# encode instance published.
207
echo "Starting prefill instance on GPU $GPU_PD, port $PREFILL_PORT"
207
- env CUDA_VISIBLE_DEVICES="$GPU_PD" VLLM_ENABLE_SCALE_OUT_ENDPOINTS=1 \
208
+ env CUDA_VISIBLE_DEVICES="$GPU_PD" \
209
vllm serve "$MODEL" \
210
+ --enable-scale-out \
211
--port "$PREFILL_PORT" \
212
--max-model-len "$MAX_MODEL_LEN" \
213
--enforce-eager \