feat: implement automatic ChromaDB indexing and loop-based documentation generation

- Add automatic ChromaDB indexing after documentation generation
- Implement loop-based section generation for individual VM and container documentation
- Fix Celery anti-pattern in generate_proxmox_docs task (removed blocking .get() call)
- Share ChromaDB vector store volume between worker and chat services
- Add documentation management UI to frontend with manual job triggering and log viewing
- Fix frontend Socket.IO connection URL to point to correct chat service port
- Enhance DocumentationAgent.index_documentation() with automatic cleanup of old documents
- Update Proxmox template to generate individual files for each VM and container

This enables the RAG system to properly respond with infrastructure-specific information from the generated documentation.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

src/datacenter_docs/utils/llm_client.py

@@ -81,7 +81,8 @@ class LLMClient:
         self.max_tokens = max_tokens or settings.LLM_MAX_TOKENS
 
         # Initialize AsyncOpenAI client with custom HTTP client (disable SSL verification for self-signed certs)
-        http_client = httpx.AsyncClient(verify=False, timeout=30.0)
+        # Increased timeout to 120s for documentation generation (large prompts)
+        http_client = httpx.AsyncClient(verify=False, timeout=120.0)
         self.client = AsyncOpenAI(
             base_url=self.base_url,
             api_key=self.api_key,
@@ -129,6 +130,13 @@ class LLMClient:
             # Type guard: we know it's ChatCompletion when stream=False
             response = cast(ChatCompletion, response)
 
+            # Check for None response or empty choices
+            if response is None:
+                raise ValueError("LLM returned None response")
+
+            if not response.choices or len(response.choices) == 0:
+                raise ValueError("LLM returned empty choices")
+
             # Extract text from first choice
             message = response.choices[0].message
             content = message.content or ""