diff --git a/config.cpu.yaml b/config.cpu.yaml
index 6ceac915..3989a547 100644
--- a/config.cpu.yaml
+++ b/config.cpu.yaml
@@ -22,10 +22,11 @@ embedding:
   # embedding service config
   # for external embedding service, set CC_EM_BASE_URL and CC_EM_APIKEY env vars during deployment
   # if the env vars are set, this config is ignored
-  # request_timeout is always respected even for remote service
+  # connect_timeout and request_timeout are always respected even for remote service
   base_url: http://localhost:5000/v1
   workers: 1
-  request_timeout: 1800 # in seconds
+  connect_timeout: 15   # in seconds
+  request_timeout: 1750 # in seconds
   # batch_size: 100 # max texts per embedding API request, 0 = no batching
   # only for external embedding service
   # remote_service: true
diff --git a/config.gpu.yaml b/config.gpu.yaml
index a12fd1be..1c9d412d 100644
--- a/config.gpu.yaml
+++ b/config.gpu.yaml
@@ -22,10 +22,11 @@ embedding:
   # embedding service config
   # for external embedding service, set CC_EM_BASE_URL and CC_EM_APIKEY env vars during deployment
   # if the env vars are set, this config is ignored
-  # request_timeout is always respected even for remote service
+  # connect_timeout and request_timeout are always respected even for remote service
   base_url: http://localhost:5000/v1
   workers: 1
-  request_timeout: 1800 # in seconds
+  connect_timeout: 15   # in seconds
+  request_timeout: 1750 # in seconds
   # batch_size: 100 # max texts per embedding API request, 0 = no batching
   # only for external embedding service
   # remote_service: true
diff --git a/context_chat_backend/config_parser.py b/context_chat_backend/config_parser.py
index 0a62019a..0ef0e2c1 100644
--- a/context_chat_backend/config_parser.py
+++ b/context_chat_backend/config_parser.py
@@ -77,7 +77,8 @@ def get_config(file_path: str) -> TConfig:
 				auth=auth,
 				remote_service=True,
 				workers=0,
-				request_timeout=embedding.get('request_timeout', 1800) if embedding else 1800,
+				connect_timeout=embedding.get('connect_timeout', 15) if embedding else 15,
+				request_timeout=embedding.get('request_timeout', 1750) if embedding else 1750,
 				batch_size=int(os.getenv('CC_EM_BATCH_SIZE', 100)),
 			)
 		except Exception as e:
diff --git a/context_chat_backend/network_em.py b/context_chat_backend/network_em.py
index 5ba8faf5..28797368 100644
--- a/context_chat_backend/network_em.py
+++ b/context_chat_backend/network_em.py
@@ -99,7 +99,10 @@ def _get_embedding(self, input_: str | list[str], try_: int = 3) -> list[float]
 			response = niquests.post(
 				f'{emconf.base_url.removesuffix("/")}/embeddings',
 				json=data,
-				timeout=emconf.request_timeout,
+				timeout=niquests.TimeoutConfiguration(
+					connect=emconf.connect_timeout,
+					read=emconf.request_timeout,
+				),
 				auth=auth,
 				verify=self.app_config.verify_ssl,
 			)
diff --git a/context_chat_backend/types.py b/context_chat_backend/types.py
index 8d7d0fcf..e85251da 100644
--- a/context_chat_backend/types.py
+++ b/context_chat_backend/types.py
@@ -104,6 +104,7 @@ class TEmbeddingAuthBasic(BaseModel):
 class TEmbeddingConfig(BaseModel):
 	base_url: str = 'http://localhost:5000/v1'
 	workers: int = 1
+	connect_timeout: int = 15
 	request_timeout: int = 1750
 	model_name: str | None = DEFAULT_EM_MODEL_ALIAS
 	auth: TEmbeddingAuthApiKey | TEmbeddingAuthBasic | None = None