add preliminary support for llm agents (#432)

demos/use_cases/orchestrating_agents/arch_config.yaml

@@ -0,0 +1,46 @@
+version: "0.1-beta"
+
+listeners:
+  ingress_traffic:
+    address: 0.0.0.0
+    port: 10000
+    message_format: openai
+    timeout: 30s
+
+  egress_traffic:
+    address: 0.0.0.0
+    port: 12000
+    message_format: openai
+    timeout: 30s
+
+overrides:
+  use_agent_orchestrator: true
+
+endpoints:
+  agent_gateway:
+    endpoint: host.docker.internal:18083
+    connect_timeout: 0.005s
+
+llm_providers:
+  - name: gpt-4o-mini
+    access_key: $OPENAI_API_KEY
+    provider_interface: openai
+    model: gpt-4o-mini
+    default: true
+
+system_prompt: |
+  You are a helpful assistant.
+
+prompt_targets:
+  - name: sales_agent
+    description: handles queries related to sales and purchases
+
+  - name: issues_and_repairs
+    description: handles issues, repairs, or refunds
+
+  - name: escalate_to_human
+    description: escalates to human agent
+
+tracing:
+  random_sampling: 100
+  trace_arch_internal: true