Migrate all callers from /get_server_info to /server_info

docs/advanced_features/server_arguments.md

@@ -212,7 +212,7 @@ Please consult the documentation below and [server_args.py](https://github.com/s
 | Argument | Description | Defaults | Options |
 | --- | --- | --- | --- |
 | `--api-key` | Set API key of the server. It is also used in the OpenAI API compatible server. | `None` | Type: str |
-| `--admin-api-key` | Set **admin API key** for administrative/control endpoints (e.g., weights update, cache flush, `/get_server_info`). Endpoints marked as admin-only require `Authorization: Bearer <admin_api_key>` when this is set. | `None` | Type: str |
+| `--admin-api-key` | Set **admin API key** for administrative/control endpoints (e.g., weights update, cache flush, `/server_info`). Endpoints marked as admin-only require `Authorization: Bearer <admin_api_key>` when this is set. | `None` | Type: str |
 | `--served-model-name` | Override the model name returned by the v1/models endpoint in OpenAI API server. | `None` | Type: str |
 | `--weight-version` | Version identifier for the model weights. Defaults to 'default' if not specified. | `default` | Type: str |
 | `--chat-template` | The builtin chat template name or the path of the chat template file. This is only used for OpenAI-compatible API server. | `None` | Type: str |

docs/advanced_features/sgl_model_gateway.md

@@ -77,7 +77,7 @@ SGLang Model Gateway is a high-performance model-routing gateway for large-scale
 
 ### Control Plane
 
-- **Worker Manager** discovers capabilities (`/get_server_info`, `/get_model_info`), tracks load, and registers/removes workers in the shared registry.
+- **Worker Manager** discovers capabilities (`/server_info`, `/get_model_info`), tracks load, and registers/removes workers in the shared registry.
 - **Job Queue** serializes add/remove requests and exposes status (`/workers/{worker_id}`) so clients can track onboarding progress.
 - **Load Monitor** feeds cache-aware and power-of-two policies with live worker load statistics.
 - **Health Checker** continuously probes workers and updates readiness, circuit breaker state, and router metrics.
@@ -552,7 +552,7 @@ Response:
 | `GET` | `/engine_metrics` | Engine-level metrics from workers |
 | `GET` | `/v1/models` | List available models |
 | `GET` | `/get_model_info` | Get model information |
-| `GET` | `/get_server_info` | Get server information |
+| `GET` | `/server_info` | Get server information |
 | `POST` | `/flush_cache` | Clear all caches |
 | `GET` | `/get_loads` | Get all worker loads |
 | `POST` | `/wasm` | Upload WASM module |

docs/basic_usage/native_api.ipynb

@@ -10,7 +10,7 @@
     "\n",
     "- `/generate` (text generation model)\n",
     "- `/get_model_info`\n",
-    "- `/get_server_info`\n",
+    "- `/server_info`\n",
     "- `/health`\n",
     "- `/health_generate`\n",
     "- `/flush_cache`\n",
@@ -140,7 +140,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "url = f\"http://localhost:{port}/get_server_info\"\n",
+    "url = f\"http://localhost:{port}/server_info\"\n",
     "\n",
     "response = requests.get(url)\n",
     "print_highlight(response.text)"

docs/developer_guide/bench_serving.md

@@ -352,4 +352,4 @@ python3 -m sglang.bench_serving \
 ### Notes
 
 - The script raises the file descriptor soft limit (`RLIMIT_NOFILE`) to help with many concurrent connections.
-- For sglang, `/get_server_info` is queried post-run to report speculative decoding accept length when available.
+- For sglang, `/server_info` is queried post-run to report speculative decoding accept length when available.

python/sglang/bench_serving.py

@@ -1402,7 +1402,7 @@ async def limited_request_func(request_func_input, pbar):
 
     if "sglang" in backend:
         server_info = requests.get(
-            base_url + "/get_server_info", headers=get_auth_headers()
+            base_url + "/server_info", headers=get_auth_headers()
         )
         if server_info.status_code == 200:
             server_info_json = server_info.json()
@@ -1538,7 +1538,7 @@ async def limited_request_func(request_func_input, pbar):
         print("{:<40} {:<10.2f}".format("Max ITL (ms):", metrics.max_itl_ms))
     print("=" * 50)
 
-    resp = requests.get(base_url + "/get_server_info", headers=get_auth_headers())
+    resp = requests.get(base_url + "/server_info", headers=get_auth_headers())
     server_info = resp.json() if resp.status_code == 200 else None
 
     if (

python/sglang/lang/backend/runtime_endpoint.py

@@ -67,7 +67,7 @@ def flush_cache(self):
 
     def get_server_info(self):
         res = http_request(
-            self.base_url + "/get_server_info",
+            self.base_url + "/server_info",
             api_key=self.api_key,
             verify=self.verify,
         )
@@ -531,7 +531,7 @@ def encode(
 
     async def get_server_info(self):
         async with aiohttp.ClientSession() as session:
-            async with session.get(f"{self.url}/get_server_info") as response:
+            async with session.get(f"{self.url}/server_info") as response:
                 if response.status == 200:
                     return await response.json()
                 else:

python/sglang/profiler.py

@@ -42,7 +42,7 @@ def run_profile(
     # Dump server args.
     file_path = Path(output_dir) / "server_args.json"
     if not file_path.exists():
-        response = requests.get(url + "/get_server_info")
+        response = requests.get(url + "/server_info")
         response.raise_for_status()
         server_args_data = response.json()
         with open(file_path, "w") as file:

python/sglang/test/bench_one_batch_server_internal.py

@@ -609,7 +609,7 @@ def run_one_case(
         last_gen_throughput = -1
         acc_length = -1
     else:
-        response = requests.get(url + "/get_server_info", timeout=DEFAULT_TIMEOUT)
+        response = requests.get(url + "/server_info", timeout=DEFAULT_TIMEOUT)
         response.raise_for_status()
         server_info = response.json()
         internal_state = server_info.get("internal_states", [{}])
@@ -793,7 +793,7 @@ def run_benchmark_internal(
         skip_max_running_requests_threshold = float("inf")
     else:
         model_name = None
-        response = requests.get(base_url + "/get_server_info", timeout=DEFAULT_TIMEOUT)
+        response = requests.get(base_url + "/server_info", timeout=DEFAULT_TIMEOUT)
         response.raise_for_status()
         server_info = response.json()
         if "tokenizer_path" in server_info:

python/sglang/test/kits/cache_hit_kit.py

@@ -221,7 +221,7 @@ async def _send_one(payload):
 def _get_page_size(base_url: str) -> int:
     """Query server for page_size used by radix cache."""
     try:
-        resp = requests.get(f"{base_url}/get_server_info", timeout=10)
+        resp = requests.get(f"{base_url}/server_info", timeout=10)
         resp.raise_for_status()
         info = resp.json()
         return info.get("page_size", 1)

python/sglang/test/kl_test_utils.py

@@ -208,7 +208,7 @@ def test_input_output_logprobs_match_helper(
 def test_input_output_logprobs_match_prefill_cache_hit_helper(
     base_url, ACC_THRESHOLDS, model_name, max_samples=None, max_new_tokens=8192
 ):
-    server_info = requests.get(base_url + "/get_server_info").json()
+    server_info = requests.get(base_url + "/server_info").json()
     if server_info["disable_radix_cache"]:
         print("Radix cache is disabled, skipping test")
         return
@@ -261,7 +261,7 @@ def test_input_output_logprobs_match_prefill_cache_hit_helper(
 def test_input_output_logprobs_match_decode_cache_hit_helper(
     base_url, ACC_THRESHOLDS, model_name, max_samples=None, max_new_tokens=8192
 ):
-    server_info = requests.get(base_url + "/get_server_info").json()
+    server_info = requests.get(base_url + "/server_info").json()
     if server_info["disable_radix_cache"]:
         print("Radix cache is disabled, skipping test")
         return

python/sglang/test/nightly_utils.py

@@ -324,7 +324,7 @@ def _get_spec_accept_length(self) -> Optional[float]:
             The average speculative decoding accept length, or None if not available.
         """
         try:
-            response = requests.get(f"{self.base_url}/get_server_info", timeout=10)
+            response = requests.get(f"{self.base_url}/server_info", timeout=10)
             if response.status_code == 200:
                 server_info = response.json()
                 internal_states = server_info.get("internal_states", [])

scripts/playground/bench_speculative.py

@@ -119,7 +119,7 @@ def send_one_batch(base_url, num_prompts, batch_size, processor, is_multimodal):
     acc_length = results["accept_length"] or 1.0
     avg_output_token = results["total_output_tokens"] / results["completed"]
 
-    server_info = requests.get(base_url + "/get_server_info").json()
+    server_info = requests.get(base_url + "/server_info").json()
     # We use 20% percentile instead of median on purpose
     step_time = np.percentile(
         server_info["internal_states"][0]["step_time_dict"][str(batch_size)], 20

sgl-model-gateway/README.md

@@ -407,7 +407,7 @@ Use upstream SGLang binaries to start dedicated worker processes.
 
 ### Worker Lifecycle & Job Queue
 - `JobQueue` handles asynchronous add/remove operations to avoid blocking clients.
-- `WorkerManager` inspects worker metadata (`/get_server_info`, `/get_model_info`), tracks load, and exposes `flush_cache` and `get_loads`.
+- `WorkerManager` inspects worker metadata (`/server_info`, `/get_model_info`), tracks load, and exposes `flush_cache` and `get_loads`.
 - Per-worker circuit breakers and health probes keep the registry healthy; load monitor feeds metrics to cache-aware and power-of-two policies.
 
 ### Administrative & Worker APIs

sgl-model-gateway/bindings/python/src/sglang_router/mini_lb.py

@@ -269,6 +269,8 @@ async def flush_cache():
     return Response(status_code=200)
 
 
+# TODO: Remove `/get_server_info` alias after one release-cycle deprecation window.
+@app.get("/server_info")
 @app.get("/get_server_info")
 async def get_server_info():
     prefill_infos = []
@@ -277,10 +279,10 @@ async def get_server_info():
 
     async with aiohttp.ClientSession() as session:
         for server in lb.prefill_urls:
-            server_info = await session.get(f"{server}/get_server_info")
+            server_info = await session.get(f"{server}/server_info")
             prefill_infos.append(await server_info.json())
         for server in lb.decode_urls:
-            server_info = await session.get(f"{server}/get_server_info")
+            server_info = await session.get(f"{server}/server_info")
             info_json = await server_info.json()
             decode_infos.append(info_json)
             # Extract internal_states from decode servers

sgl-model-gateway/src/routers/http/pd_router.rs

@@ -1223,7 +1223,7 @@ impl RouterTrait for PDRouter {
     async fn get_server_info(&self, _req: Request<Body>) -> Response {
         // Get info from the first decode server to match sglang's server info format
         // Note: We use decode workers for server info to match expected format
-        self.proxy_to_first_prefill_worker("get_server_info", None)
+        self.proxy_to_first_prefill_worker("server_info", None)
             .await
     }
 

sgl-model-gateway/src/routers/http/router.rs

@@ -724,7 +724,7 @@ impl RouterTrait for Router {
     }
 
     async fn get_server_info(&self, req: Request<Body>) -> Response {
-        self.proxy_get_request(req, "get_server_info").await
+        self.proxy_get_request(req, "server_info").await
     }
 
     async fn get_models(&self, req: Request<Body>) -> Response {

sgl-model-gateway/src/server.rs

@@ -610,6 +610,8 @@ pub fn build_app(
         .route("/engine_metrics", get(engine_metrics))
         .route("/v1/models", get(v1_models))
         .route("/get_model_info", get(get_model_info))
+        .route("/server_info", get(get_server_info))
+        // TODO: Remove `/get_server_info` alias after one release-cycle deprecation window.
         .route("/get_server_info", get(get_server_info));
 
     // Build admin routes with control plane auth if configured, otherwise use simple API key auth

sgl-model-gateway/tests/api/api_endpoints_test.rs

@@ -314,7 +314,7 @@ mod model_info_tests {
 
         let req = Request::builder()
             .method("GET")
-            .uri("/get_server_info")
+            .uri("/server_info")
             .body(Body::empty())
             .unwrap();
 
@@ -445,7 +445,7 @@ mod model_info_tests {
 
         let req = Request::builder()
             .method("GET")
-            .uri("/get_server_info")
+            .uri("/server_info")
             .body(Body::empty())
             .unwrap();
         let resp = app.clone().oneshot(req).await.unwrap();

sgl-model-gateway/tests/common/mock_worker.rs

@@ -82,7 +82,7 @@ impl MockWorker {
         let app = Router::new()
             .route("/health", get(health_handler))
             .route("/health_generate", get(health_generate_handler))
-            .route("/get_server_info", get(server_info_handler))
+            .route("/server_info", get(server_info_handler))
             .route("/get_model_info", get(model_info_handler))
             .route("/generate", post(generate_handler))
             .route("/v1/chat/completions", post(chat_completions_handler))

sgl-model-gateway/tests/common/tls_mock_worker.rs

@@ -101,7 +101,7 @@ impl TlsMockWorker {
         let app = Router::new()
             .route("/health", get(health_handler))
             .route("/health_generate", get(health_generate_handler))
-            .route("/get_server_info", get(server_info_handler))
+            .route("/server_info", get(server_info_handler))
             .route("/generate", post(generate_handler))
             .route("/v1/chat/completions", post(chat_completions_handler))
             .with_state(config);

sgl-model-gateway/tests/routing/test_pd_routing.rs

@@ -765,7 +765,7 @@ mod pd_routing_unit_tests {
         let implemented_endpoints = vec![
             ("/health", "GET", true),
             ("/health_generate", "GET", true), // Note: Python uses POST, we use GET
-            ("/get_server_info", "GET", true),
+            ("/server_info", "GET", true),
             ("/v1/models", "GET", true),
             ("/get_model_info", "GET", true),
             ("/generate", "POST", true),

test/registered/4-gpu-models/test_qwen35_models.py

@@ -149,7 +149,7 @@ def test_gsm8k(self):
         print(f"{metrics=}")
         self.assertGreaterEqual(metrics["score"], ACC_THRESHOLDS[self.model]["gsm8k"])
 
-        server_info = requests.get(self.base_url + "/get_server_info")
+        server_info = requests.get(self.base_url + "/server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]
@@ -226,7 +226,7 @@ def test_gsm8k(self):
         print(f"{metrics=}")
         self.assertGreaterEqual(metrics["score"], ACC_THRESHOLDS[self.model]["gsm8k"])
 
-        server_info = requests.get(self.base_url + "/get_server_info")
+        server_info = requests.get(self.base_url + "/server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/registered/8-gpu-models/test_deepseek_v32_mtp.py

@@ -76,7 +76,7 @@ def test_a_gsm8k(
         metrics = run_eval(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/get_server_info")
+        server_info = requests.get(self.base_url + "/server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]
@@ -163,7 +163,7 @@ def test_a_gsm8k(
         metrics = run_eval(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/get_server_info")
+        server_info = requests.get(self.base_url + "/server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]
@@ -246,7 +246,7 @@ def test_a_gsm8k(
         metrics = run_eval(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/get_server_info")
+        server_info = requests.get(self.base_url + "/server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]
@@ -330,7 +330,7 @@ def test_a_gsm8k(
         metrics = run_eval(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/get_server_info")
+        server_info = requests.get(self.base_url + "/server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/registered/8-gpu-models/test_deepseek_v3_mtp.py

@@ -72,7 +72,7 @@ def test_a_gsm8k(
         metrics = run_eval(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/get_server_info")
+        server_info = requests.get(self.base_url + "/server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/registered/amd/accuracy/mi30x/test_deepseek_v32_mtp_eval_amd.py

@@ -106,7 +106,7 @@ def test_a_gsm8k(self):
         metrics = run_eval_few_shot_gsm8k(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/get_server_info")
+        server_info = requests.get(self.base_url + "/server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/registered/amd/accuracy/mi35x/test_deepseek_v32_mtp_eval_mi35x.py

@@ -108,7 +108,7 @@ def test_a_gsm8k(self):
         metrics = run_eval_few_shot_gsm8k(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/get_server_info")
+        server_info = requests.get(self.base_url + "/server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/registered/amd/test_deepseek_r1_mxfp4_8gpu.py

@@ -135,7 +135,7 @@ def test_a_gsm8k(
         metrics = run_eval_few_shot_gsm8k(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/get_server_info")
+        server_info = requests.get(self.base_url + "/server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/registered/amd/test_deepseek_v32_mtp.py

@@ -87,7 +87,7 @@ def test_a_gsm8k(
         metrics = run_eval_few_shot_gsm8k(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/get_server_info")
+        server_info = requests.get(self.base_url + "/server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]
@@ -179,7 +179,7 @@ def test_a_gsm8k(
         metrics = run_eval_few_shot_gsm8k(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/get_server_info")
+        server_info = requests.get(self.base_url + "/server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/registered/amd/test_deepseek_v3_mtp.py

@@ -72,7 +72,7 @@ def test_a_gsm8k(
         metrics = run_eval_few_shot_gsm8k(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/get_server_info")
+        server_info = requests.get(self.base_url + "/server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/registered/amd/test_deepseek_v3_mtp_kv_fp8.py

@@ -76,7 +76,7 @@ def test_a_gsm8k(
         metrics = run_eval_few_shot_gsm8k(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/get_server_info")
+        server_info = requests.get(self.base_url + "/server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]