fix

arch/arch_config_schema.yaml

@@ -33,6 +33,8 @@ properties:
             enum:
               - http
               - https
+          http_host:
+            type: string
         additionalProperties: false
         required:
           - endpoint

arch/envoy.template.yaml

@@ -486,14 +486,14 @@ static_resources:
         typed_config:
           "@type": type.googleapis.com/envoy.extensions.transport_sockets.tls.v3.UpstreamTlsContext
           sni: api.mistral.ai
-    {% for internal_clustrer in ["arch_fc", "model_server"] %}
-    - name: {{ internal_clustrer }}
+    {% for internal_cluster in ["arch_fc", "model_server"] %}
+    - name: {{ internal_cluster }}
       connect_timeout: 5s
       type: STRICT_DNS
       dns_lookup_family: V4_ONLY
       lb_policy: ROUND_ROBIN
       load_assignment:
-        cluster_name: {{ internal_clustrer }}
+        cluster_name: {{ internal_cluster }}
         endpoints:
           - lb_endpoints:
               - endpoint:
@@ -501,7 +501,7 @@ static_resources:
                     socket_address:
                       address: host.docker.internal
                       port_value: $MODEL_SERVER_PORT
-                  hostname: {{ internal_clustrer }}
+                  hostname: {{ internal_cluster }}
     {% endfor %}
     - name: mistral_7b_instruct
       connect_timeout: 5s
@@ -537,7 +537,11 @@ static_resources:
                     socket_address:
                       address: {{ cluster.endpoint }}
                       port_value: {{ cluster.port }}
+                  {% if cluster.http_host %}
+                  hostname: {{ cluster.http_host }}
+                  {% else %}
                   hostname: {{ cluster.endpoint }}
+                  {% endif %}
       {% if cluster.protocol == "https" %}
       transport_socket:
         name: envoy.transport_sockets.tls

arch/tools/cli/utils.py

@@ -89,6 +89,18 @@ def get_llm_provider_access_keys(arch_config_file):
         if acess_key is not None:
             access_key_list.append(acess_key)
 
+    for prompt_target in arch_config_yaml.get("prompt_targets", []):
+        for k, v in prompt_target.get("endpoint", {}).get("http_headers", {}).items():
+            if k.lower() == "authorization":
+                print(
+                    f"found auth header: {k} for prompt_target: {prompt_target.get('name')}/{prompt_target.get('endpoint').get('name')}"
+                )
+                auth_tokens = v.split(" ")
+                if len(auth_tokens) > 1:
+                    access_key_list.append(auth_tokens[1])
+                else:
+                    access_key_list.append(v)
+
     return access_key_list