Update configurations for HF

Dockerfile

@@ -5,6 +5,7 @@ WORKDIR /app
 # Install system dependencies
 RUN apt-get update && apt-get install -y \
     build-essential \
+    curl \
     && rm -rf /var/lib/apt/lists/*
 
 # Copy requirements first to leverage Docker cache

docker-compose.yml

@@ -3,14 +3,15 @@ version: '3.8'
 services:
   web:
     build: .
-    command: gunicorn --config gunicorn_config.py app:app --reload
+    command: gunicorn --config gunicorn_config.py app:app
     expose:
       - "8008"
     volumes:
       - .:/app
       - upload_data:/app/uploads
     depends_on:
-      - redis
+      redis:
+        condition: service_healthy
     environment:
       - FLASK_ENV=production
       - REDIS_URL=redis://redis:6379/0
@@ -19,7 +20,13 @@ services:
       resources:
         limits:
           cpus: '0.6'
-          memory: 4G
+          memory: 6G
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:8008/health"]
+      interval: 30s
+      timeout: 10s
+      retries: 3
+      start_period: 40s
     restart: unless-stopped
 
   redis:
@@ -29,6 +36,11 @@ services:
       - redis_data:/data
     ports:
       - "6379:6379"
+    healthcheck:
+      test: ["CMD", "redis-cli", "ping"]
+      interval: 10s
+      timeout: 5s
+      retries: 3
     deploy:
       resources:
         limits:

gunicorn_config.py

@@ -1,17 +1,17 @@
 import multiprocessing
 import os
 
-# Number of worker processes - adjust based on CPU cores
-workers = multiprocessing.cpu_count() * 2 + 1
+# Number of worker processes - limited for 2vCPU environment
+workers = 2  # Using 2 workers for 2vCPU
 
-# Number of threads per worker
-threads = 4
+# Number of threads per worker - reduced for memory efficiency
+threads = 2
 
 # Maximum number of pending connections
-backlog = 2048
+backlog = 1024
 
 # Maximum number of requests a worker will process before restarting
-max_requests = 10000
+max_requests = 1000
 max_requests_jitter = 50
 
 # Timeout for worker processes (5 minutes)
@@ -30,8 +30,8 @@ access_log_format = '%(h)s %(l)s %(u)s %(t)s "%(r)s" %(s)s %(b)s "%(f)s" "%(a)s"
 port = os.getenv('PORT', '8008')  # HF uses PORT env var
 bind = f"0.0.0.0:{port}"
 
-# Worker class
-worker_class = "gevent"
+# Worker class - using sync for better stability
+worker_class = "sync"
 
 # Process name
 proc_name = "selector_server"