fix: redis url not properly obtained

.env.ci

@@ -42,8 +42,7 @@ POSTGRES_PORT = 5432
 REDIS_URL = "redis://redis:6379"
 
 # Model Discovery Redis Docker Compose Config
-DISCOVERY_HOST = "redis"
-DISCOVERY_PORT = 6379
+DISCOVERY_URL = "redis://redis:6379"
 
 # Grafana Docker Compose Config
 GF_SECURITY_ADMIN_USER = "admin"

.env.sample

@@ -37,6 +37,9 @@ POSTGRES_PORT = 5432
 # Redis Docker Compose Config
 REDIS_URL = "redis://redis:6379"
 
+# Model Discovery Redis Docker Compose Config
+DISCOVERY_URL = "redis://redis:6379"
+
 # Etcd Docker Compose Config
 ETCD_HOST = "etcd"
 ETCD_PORT = 2379

docker/compose/docker-compose.deepseek-14b-gpu.yml

@@ -24,8 +24,7 @@ services:
     environment:
       - SVC_HOST=deepseek_14b_gpu
       - SVC_PORT=8000
-      - DISCOVERY_HOST=redis
-      - DISCOVERY_PORT=6379
+      - DISCOVERY_URL=redis://redis:6379
       - TOOL_SUPPORT=false
     volumes:
       - hugging_face_models:/root/.cache/huggingface  # cache models

docker/compose/docker-compose.gemma-27b-gpu.yml

@@ -29,8 +29,7 @@ services:
     environment:
       - SVC_HOST=gemma_27b_gpu
       - SVC_PORT=8000
-      - DISCOVERY_HOST=redis
-      - DISCOVERY_PORT=6379
+      - DISCOVERY_URL=redis://redis:6379
       - TOOL_SUPPORT=false
       - MULTIMODAL_SUPPORT=true
       - MODEL_NUM_RETRIES=60

docker/compose/docker-compose.gemma-4b-gpu.ci.yml

@@ -28,8 +28,7 @@ services:
     environment:
       - SVC_HOST=gemma_4b_gpu
       - SVC_PORT=8000
-      - DISCOVERY_HOST=redis
-      - DISCOVERY_PORT=6379
+      - DISCOVERY_URL=redis://redis:6379
       - TOOL_SUPPORT=false
       - MULTIMODAL_SUPPORT=true
       - CUDA_LAUNCH_BLOCKING=1

docker/compose/docker-compose.gpt-120b-gpu.yml

@@ -28,8 +28,7 @@ services:
     environment:
       - SVC_HOST=gpt_120b_gpu
       - SVC_PORT=8000
-      - DISCOVERY_HOST=redis
-      - DISCOVERY_PORT=6379
+      - DISCOVERY_URL=redis://redis:6379
       - TOOL_SUPPORT=true
     volumes:
       - hugging_face_models:/root/.cache/huggingface  # cache models

docker/compose/docker-compose.gpt-20b-gpu.ci.yml

@@ -23,8 +23,7 @@ services:
     environment:
       - SVC_HOST=gpt_20b_gpu
       - SVC_PORT=8000
-      - DISCOVERY_HOST=redis
-      - DISCOVERY_PORT=6379
+      - DISCOVERY_URL=redis://redis:6379
       - TOOL_SUPPORT=true
     volumes:
       - hugging_face_models:/root/.cache/huggingface # cache models

docker/compose/docker-compose.gpt-20b-gpu.yml

@@ -28,8 +28,7 @@ services:
     environment:
       - SVC_HOST=gpt_20b_gpu
       - SVC_PORT=8000
-      - DISCOVERY_HOST=redis
-      - DISCOVERY_PORT=6379
+      - DISCOVERY_URL=redis://redis:6379
       - TOOL_SUPPORT=true
     volumes:
       - hugging_face_models:/root/.cache/huggingface  # cache models

docker/compose/docker-compose.llama-1b-cpu.yml

@@ -19,8 +19,7 @@ services:
     environment:
       - SVC_HOST=llama_1b_cpu
       - SVC_PORT=8000
-      - DISCOVERY_HOST=redis
-      - DISCOVERY_PORT=6379
+      - DISCOVERY_URL=redis://redis:6379
       - TOOL_SUPPORT=true
     volumes:
       - hugging_face_models:/root/.cache/huggingface

docker/compose/docker-compose.llama-1b-gpu.ci.yml

@@ -32,8 +32,7 @@ services:
     environment:
       - SVC_HOST=llama_1b_gpu
       - SVC_PORT=8000
-      - DISCOVERY_HOST=redis
-      - DISCOVERY_PORT=6379
+      - DISCOVERY_URL=redis://redis:6379
       - TOOL_SUPPORT=true
       - CUDA_LAUNCH_BLOCKING=1
     volumes:

docker/compose/docker-compose.llama-1b-gpu.yml

@@ -30,8 +30,7 @@ services:
     environment:
       - SVC_HOST=llama_1b_gpu
       - SVC_PORT=8000
-      - DISCOVERY_HOST=redis
-      - DISCOVERY_PORT=6379
+      - DISCOVERY_URL=redis://redis:6379
       - TOOL_SUPPORT=true
     volumes:
       - hugging_face_models:/root/.cache/huggingface  # cache models

docker/compose/docker-compose.llama-3b-gpu.yml

@@ -30,8 +30,7 @@ services:
     environment:
       - SVC_HOST=llama_3b_gpu
       - SVC_PORT=8000
-      - DISCOVERY_HOST=redis
-      - DISCOVERY_PORT=6379
+      - DISCOVERY_URL=redis://redis:6379
       - TOOL_SUPPORT=true
     volumes:
       - hugging_face_models:/root/.cache/huggingface  # cache models

docker/compose/docker-compose.llama-70b-gpu.yml

@@ -30,8 +30,7 @@ services:
     environment:
       - SVC_HOST=llama_70b_gpu
       - SVC_PORT=8000
-      - DISCOVERY_HOST=redis
-      - DISCOVERY_PORT=6379
+      - DISCOVERY_URL=redis://redis:6379
       - TOOL_SUPPORT=true
     volumes:
       - hugging_face_models:/root/.cache/huggingface  # cache models

docker/compose/docker-compose.llama-8b-gpu.yml

@@ -31,8 +31,7 @@ services:
     environment:
       - SVC_HOST=llama_8b_gpu
       - SVC_PORT=8000
-      - DISCOVERY_HOST=redis
-      - DISCOVERY_PORT=6379
+      - DISCOVERY_URL=redis://redis:6379
       - TOOL_SUPPORT=true
     volumes:
       - hugging_face_models:/root/.cache/huggingface

docker/compose/docker-compose.lmstudio.yml

@@ -9,8 +9,7 @@ services:
     environment:
       - SVC_HOST=host.docker.internal
       - SVC_PORT=1234
-      - DISCOVERY_HOST=redis
-      - DISCOVERY_PORT=6379
+      - DISCOVERY_URL=redis://redis:6379
       - LMSTUDIO_SUPPORTED_FEATURES=chat_completion
     extra_hosts:
       - "host.docker.internal:host-gateway"

docker/compose/docker-compose.nilai-prod-1.yml

@@ -29,8 +29,7 @@ services:
     environment:
       - SVC_HOST=gemma_27b_gpu
       - SVC_PORT=8000
-      - DISCOVERY_HOST=redis
-      - DISCOVERY_PORT=6379
+      - DISCOVERY_URL=redis://redis:6379
       - TOOL_SUPPORT=false
       - MULTIMODAL_SUPPORT=true
       - MODEL_NUM_RETRIES=60

docker/compose/docker-compose.nilai-prod-2.yml

@@ -37,8 +37,7 @@ services:
     environment:
       - SVC_HOST=llama_8b_gpu
       - SVC_PORT=8000
-      - DISCOVERY_HOST=redis
-      - DISCOVERY_PORT=6379
+      - DISCOVERY_URL=redis://redis:6379
       - TOOL_SUPPORT=true
     volumes:
       - hugging_face_models:/root/.cache/huggingface
@@ -77,8 +76,7 @@ services:
     environment:
       - SVC_HOST=gpt_20b_gpu
       - SVC_PORT=8000
-      - DISCOVERY_HOST=redis
-      - DISCOVERY_PORT=6379
+      - DISCOVERY_URL=redis://redis:6379
       - TOOL_SUPPORT=true
     volumes:
       - hugging_face_models:/root/.cache/huggingface  # cache models

docker/compose/docker-compose.qwen-2b-gpu.ci.yml

@@ -44,8 +44,7 @@ services:
     environment:
       SVC_HOST: qwen2vl_2b_gpu
       SVC_PORT: "8000"
-      DISCOVERY_HOST: redis
-      DISCOVERY_PORT: "6379"
+      DISCOVERY_URL: redis://redis:6379
       TOOL_SUPPORT: "true"
       MULTIMODAL_SUPPORT: "true"
       CUDA_LAUNCH_BLOCKING: "1"

nilai-api/src/nilai_api/config/database.py

@@ -10,8 +10,10 @@ class DatabaseConfig(BaseModel):
 
 
 class DiscoveryConfig(BaseModel):
-    host: str = Field(default="localhost", description="Redis host for discovery")
-    port: int = Field(default=6379, description="Redis port for discovery")
+    url: str = Field(
+        default="redis://localhost:6379",
+        description="Redis URL for discovery (preferred default)",
+    )
 
 
 class RedisConfig(BaseModel):

nilai-api/src/nilai_api/state.py

@@ -16,9 +16,7 @@ def __init__(self):
         self.private_key, self.public_key, self.b64_public_key = generate_key_pair()
         self.sem = Semaphore(2)
 
-        self.discovery_service = ModelServiceDiscovery(
-            host=CONFIG.discovery.host, port=CONFIG.discovery.port
-        )
+        self.discovery_service = ModelServiceDiscovery(url=CONFIG.discovery.url)
         self._discovery_initialized = False
         self._uptime = time.time()
 

nilai-models/src/nilai_models/daemon.py

@@ -89,9 +89,7 @@ async def main():
     logging.basicConfig(level=logging.INFO)
 
     # Initialize discovery service
-    discovery_service = ModelServiceDiscovery(
-        host=SETTINGS.discovery_host, port=SETTINGS.discovery_port
-    )
+    discovery_service = ModelServiceDiscovery(url=SETTINGS.discovery_url)
     await discovery_service.initialize()
 
     # Fetch metadata and create endpoint

nilai-models/src/nilai_models/lmstudio_announcer.py

@@ -76,15 +76,12 @@ async def _fetch_model_ids(
 async def _announce_model(
     metadata: ModelMetadata,
     base_url: str,
-    discovery_host: str,
-    discovery_port: int,
+    discovery_url: str,
     lease_ttl: int,
     prefix: str,
 ):
     """Register and maintain a model announcement in Redis."""
-    discovery = ModelServiceDiscovery(
-        host=discovery_host, port=discovery_port, lease_ttl=lease_ttl
-    )
+    discovery = ModelServiceDiscovery(url=discovery_url, lease_ttl=lease_ttl)
     await discovery.initialize()
 
     endpoint = ModelEndpoint(url=base_url.rstrip("/"), metadata=metadata)
@@ -198,8 +195,7 @@ async def main():
         "Announcing LMStudio models %s via %s with Redis at %s:%s",
         ", ".join(model_ids),
         registration_url,
-        SETTINGS.discovery_host,
-        SETTINGS.discovery_port,
+        SETTINGS.discovery_url,
     )
 
     # Create announcement tasks for all models
@@ -220,8 +216,7 @@ async def main():
                     multimodal_default,
                 ),
                 base_url=registration_url,
-                discovery_host=SETTINGS.discovery_host,
-                discovery_port=SETTINGS.discovery_port,
+                discovery_url=SETTINGS.discovery_url,
                 lease_ttl=lease_ttl,
                 prefix=discovery_prefix,
             )

packages/nilai-common/src/nilai_common/config/host.py

@@ -14,11 +14,9 @@ class HostSettings(BaseModel):
 
     host: str = Field(default="localhost", description="Host of the service")
     port: int = Field(default=8000, description="Port of the service")
-    discovery_host: str = Field(
-        default="localhost", description="Host of the discovery service"
-    )
-    discovery_port: int = Field(
-        default=6379, description="Port of the discovery service"
+    discovery_url: str = Field(
+        default="redis://redis:6379",
+        description="Redis URL of the discovery service (preferred)",
     )
     gunicorn_workers: int = Field(default=10, description="Number of gunicorn workers")
 
@@ -27,7 +25,6 @@ class HostSettings(BaseModel):
 SETTINGS: HostSettings = HostSettings(
     host=str(os.getenv("SVC_HOST", "localhost")),
     port=int(os.getenv("SVC_PORT", 8000)),
-    discovery_host=str(os.getenv("DISCOVERY_HOST", "redis")),
-    discovery_port=int(os.getenv("DISCOVERY_PORT", 6379)),
+    discovery_url=str(os.getenv("DISCOVERY_URL", "redis://redis:6379")),
     gunicorn_workers=int(os.getenv("NILAI_GUNICORN_WORKERS", 10)),
 )

packages/nilai-common/src/nilai_common/discovery.py

@@ -14,17 +14,21 @@
 
 
 class ModelServiceDiscovery:
-    def __init__(self, host: str = "localhost", port: int = 6379, lease_ttl: int = 60):
+    def __init__(
+        self,
+        url: str = "redis://localhost:6379",
+        lease_ttl: int = 60,
+    ):
         """
         Initialize Redis client for model service discovery.
 
+        :param url: Redis URL (e.g., redis:// or rediss://). Preferred default.
         :param host: Redis server host
         :param port: Redis server port
         :param lease_ttl: TTL time for endpoint registration (in seconds)
         """
-        self.host = host
-        self.port = port
         self.lease_ttl = lease_ttl
+        self.url = url
         self._client: Optional[redis.Redis] = None
         self._model_key: Optional[str] = None
 
@@ -39,9 +43,10 @@ async def initialize(self):
         Initialize the Redis client.
         """
         if self._client is None:
-            self._client = await redis.Redis(
-                host=self.host, port=self.port, decode_responses=True
-            )
+            if self.url:
+                self._client = redis.from_url(self.url, decode_responses=True)
+            else:
+                raise ValueError("A Redis URL is required for service discovery")
 
     @property
     async def client(self) -> redis.Redis:

tests/unit/nilai-common/test_discovery.py

@@ -10,7 +10,7 @@
 async def model_service_discovery(redis_host_port):
     """Create a ModelServiceDiscovery instance connected to the test Redis container."""
     host, port = redis_host_port
-    discovery = ModelServiceDiscovery(host=host, port=port, lease_ttl=60)
+    discovery = ModelServiceDiscovery(url=f"redis://{host}:{port}", lease_ttl=60)
     await discovery.initialize()
     yield discovery
     await discovery.close()
@@ -172,8 +172,7 @@ async def test_keep_alive(model_service_discovery, model_endpoint):
     """Test the keep_alive functionality that refreshes TTL."""
     # Register a model with a short TTL
     short_ttl_discovery = ModelServiceDiscovery(
-        host=model_service_discovery.host,
-        port=model_service_discovery.port,
+        url=model_service_discovery.url,
         lease_ttl=2,  # 2 second TTL
     )
     await short_ttl_discovery.initialize()
@@ -214,8 +213,7 @@ async def test_keep_alive_with_stored_key(model_service_discovery, model_endpoin
     """Test keep_alive using the stored key from registration."""
     # Register a model with a short TTL
     short_ttl_discovery = ModelServiceDiscovery(
-        host=model_service_discovery.host,
-        port=model_service_discovery.port,
+        url=model_service_discovery.url,
         lease_ttl=2,  # 2 second TTL
     )
     await short_ttl_discovery.initialize()