archgw: address drift in prometheus cluster name

inference-platforms/archgw/README.md

@@ -64,6 +64,9 @@ and anything added in Arch Gateway's [wasm filter][archgw-wasm].
   instructions to run from Docker (to avoid nested docker).
 * Traces come from Envoy, whose configuration is written by `archgw`. At the
   moment, this hard-codes aspects including default ports.
+* Prometheus metrics show the cluster as "openai_host" - the provider_interface
+  plus the first segment of the hostname (dots truncate the rest). The "host"
+  comes from "host.docker.internal".
 * Until [this][openai-responses] resolves, don't use `--use-responses-api`.
 
 The chat prompt was designed to be idempotent, but the results are not. You may

inference-platforms/archgw/arch_config.yaml

@@ -8,7 +8,8 @@ listeners:
     timeout: 30s
 
 llm_providers:
-  - model: ollama/qwen3:0.6b
+  # We don't use 'ollama', as we want to use 'openai' provider_interface.
+  - model: local/qwen3:0.6b
     provider_interface: openai
     # This configuration is converted to Envoy and run inside Docker.
     base_url: http://host.docker.internal:11434

inference-platforms/archgw/docker-compose-elastic.yml

@@ -2,8 +2,7 @@ configs:
   # Configuration is simplified from archgw here:
   # https://github.com/katanemo/archgw/blob/main/docs/source/guides/observability/monitoring.rst
   #
-  # Note: The prometheus cluster name for qwen3:0.65b will shows up as '6b'
-  # See https://github.com/katanemo/archgw/issues/504
+  # Note: The cluster name for openai + host.docker.internal = openai_host
   prometheus-pump-config:
     content: |
       receivers: