fix(infra): use bash socket healthcheck for qdrant in staging and test (#68)

This Pull Request fixes the Qdrant startup error on the Staging and Testing environments.

### 🔍 Cause of the Bug
The official `qdrant/qdrant:latest` image is built on `debian-slim` and **does not contain `curl` or `wget`**. Changing the healthcheck to `curl` caused Qdrant to exit with status `127` (command not found), marking the service as unhealthy/error in Docker.

### 🛠️ Solution
Reverts the healthcheck in both `docker-compose.stage.yml` and `docker-compose.test.yml` to the robust, built-in bash TCP socket check:
```yaml
    healthcheck:
      test: ["CMD-SHELL", "bash -c 'exec 3<>/dev/tcp/127.0.0.1/6333'"]
```

Successfully validated locally and tested compilation.

---------

Co-authored-by: Marek Jasiński <jasins.marek@gmail.com>
Reviewed-on: #68
Co-authored-by: Antigravity <antigravity@google.com>
Co-committed-by: Antigravity <antigravity@google.com>

docker-compose.test.yml

@@ -59,7 +59,7 @@ services:
     volumes:
       - qdrant_test_data:/qdrant/storage
     healthcheck:
-      test: ["CMD-SHELL", "curl -sf http://localhost:6333/healthz || exit 1"]
+      test: ["CMD-SHELL", "bash -c 'exec 3<>/dev/tcp/127.0.0.1/6333'"]
       interval: 5s
       timeout: 5s
       retries: 5