first commit

examples/dockers/data/config.json

@@ -0,0 +1,40 @@
+{
+  "$schema": "https://www.getbifrost.ai/schema",
+  "config_store": {
+    "enabled": true,
+    "type": "sqlite",
+    "config": {
+      "path": "/app/data/config.db"
+    }
+  },
+  "logs_store": {
+    "enabled": true,
+    "type": "sqlite",
+    "config": {
+      "path": "/app/data/logs.db"
+    }
+  },
+  "client": {
+    "drop_excess_requests": false,
+    "initial_pool_size": 5000,
+    "prometheus_labels": [],
+    "enable_logging": true,
+    "disable_content_logging": false,
+    "log_retention_days": 365,
+    "enforce_auth_on_inference": false,
+    "allow_direct_keys": false,
+    "allowed_origins": [
+      "*"
+    ],
+    "max_request_body_size_mb": 100,
+    "compat": {
+      "should_convert_params": false
+    }
+  },
+  "framework": {
+    "pricing": {
+      "pricing_url": "https://getbifrost.ai/datasheet",
+      "pricing_sync_interval": 86400
+    }
+  }
+}

examples/dockers/docker-compose.yml

@@ -0,0 +1,36 @@
+services:
+  bifrost:
+    image: maximhq/bifrost:v1.4.3
+    container_name: bifrost
+    ports:
+      - "8080:8080"
+    volumes:
+      - ./data:/app/data
+    environment:
+      - APP_PORT=8080
+      - APP_HOST=0.0.0.0
+      - LOG_LEVEL=info
+      - LOG_STYLE=json
+      # Go runtime performance tuning (uncomment and adjust for your workload):
+      # - GOGC=200                  # Higher = less GC, more memory (default: 100)
+      # - GOMEMLIMIT=3600MiB        # Set to ~90% of container memory limit
+    # File descriptor limits for high-concurrency workloads
+    ulimits:
+      nofile:
+        soft: 65536
+        hard: 65536
+    # Resource limits (uncomment and adjust based on your infrastructure)
+    # deploy:
+    #   resources:
+    #     limits:
+    #       cpus: '4'
+    #       memory: 4G
+    #     reservations:
+    #       cpus: '2'
+    #       memory: 2G
+    healthcheck:
+      test: ["CMD", "wget", "--no-verbose", "--tries=1", "-O", "/dev/null", "http://localhost:8080/health"]
+      interval: 30s
+      timeout: 10s
+      retries: 3
+    restart: unless-stopped