fix: Add the option to not verify SSL at remote-vllm provider (#1585)

# What does this PR do?
Add the option to not verify SSL certificates for the remote-vllm
provider. This allows llama stack server to talk to remote LLMs which
have self-signed certificates

Partially addresses  #1545

llama_stack/providers/remote/inference/vllm/config.py

@@ -25,6 +25,10 @@ class VLLMInferenceAdapterConfig(BaseModel):
         default="fake",
         description="The API token",
     )
+    tls_verify: bool = Field(
+        default=True,
+        description="Whether to verify TLS certificates",
+    )
 
     @classmethod
     def sample_run_config(
@@ -36,4 +40,5 @@ class VLLMInferenceAdapterConfig(BaseModel):
             "url": url,
             "max_tokens": "${env.VLLM_MAX_TOKENS:4096}",
             "api_token": "${env.VLLM_API_TOKEN:fake}",
+            "tls_verify": "${env.VLLM_TLS_VERIFY:true}",
         }

llama_stack/providers/remote/inference/vllm/vllm.py

@@ -7,6 +7,7 @@ import json
 import logging
 from typing import AsyncGenerator, List, Optional, Union
 
+import httpx
 from openai import AsyncOpenAI
 from openai.types.chat.chat_completion_chunk import (
     ChatCompletionChunk as OpenAIChatCompletionChunk,
@@ -229,7 +230,11 @@ class VLLMInferenceAdapter(Inference, ModelsProtocolPrivate):
 
     async def initialize(self) -> None:
         log.info(f"Initializing VLLM client with base_url={self.config.url}")
-        self.client = AsyncOpenAI(base_url=self.config.url, api_key=self.config.api_token)
+        self.client = AsyncOpenAI(
+            base_url=self.config.url,
+            api_key=self.config.api_token,
+            http_client=None if self.config.tls_verify else httpx.AsyncClient(verify=False),
+        )
 
     async def shutdown(self) -> None:
         pass

llama_stack/templates/remote-vllm/run-with-safety.yaml

@@ -18,12 +18,14 @@ providers:
       url: ${env.VLLM_URL}
       max_tokens: ${env.VLLM_MAX_TOKENS:4096}
       api_token: ${env.VLLM_API_TOKEN:fake}
+      tls_verify: ${env.VLLM_TLS_VERIFY:true}
   - provider_id: vllm-safety
     provider_type: remote::vllm
     config:
       url: ${env.SAFETY_VLLM_URL}
       max_tokens: ${env.VLLM_MAX_TOKENS:4096}
       api_token: ${env.VLLM_API_TOKEN:fake}
+      tls_verify: ${env.VLLM_TLS_VERIFY:true}
   - provider_id: sentence-transformers
     provider_type: inline::sentence-transformers
     config: {}

llama_stack/templates/remote-vllm/run.yaml

@@ -18,6 +18,7 @@ providers:
       url: ${env.VLLM_URL}
       max_tokens: ${env.VLLM_MAX_TOKENS:4096}
       api_token: ${env.VLLM_API_TOKEN:fake}
+      tls_verify: ${env.VLLM_TLS_VERIFY:true}
   - provider_id: sentence-transformers
     provider_type: inline::sentence-transformers
     config: {}