Remove the Grafana MCP server

Reverts the declaration. The approach does not hold up in practice: an MCP
client resends the whole tool list every turn, so simply having this server
connected cost ~19k prompt tokens per request — and ~5.4k even after
trimming the tool set by 72%. On top of that its metric-name listing
returns 3330 entries, 137 KiB, which the assistant called twice while
exploring. Conversations tripped Kong's 50k-tokens-per-minute AI rate limit
and sat in a 429 retry loop.

The deployment, the Vault-backed secret and the Grafana service account
stay in place; only the LibreChat wiring goes.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

k8s/librechat/configmap-app.yaml

@@ -14,23 +14,7 @@ data:
         - "https://sales-dev.sttlab.local"
         - "https://workforce-dev.sttlab.local"
         - "https://keycloak.sttlab.local"
-        # Cluster-internal and unencrypted, unlike every other entry: the
-        # Grafana MCP server is reached directly, not through a gateway.
-        - "http://mcp-grafana.observability.svc.cluster.local:8000"
     mcpServers:
-      # Grafana's own MCP server: Prometheus, Loki and Tempo as tools, so the
-      # telemetry can be questioned instead of browsed.
-      #
-      # No oauth block, unlike every other server here. The others sit behind a
-      # Kong gateway enforcing openid-connect; this one is reached directly on
-      # the cluster network and authenticates to Grafana with a service account
-      # token of its own. LibreChat passes no user identity along, so there is
-      # no per-user scoping either — every LibreChat user gets the same full
-      # read of all telemetry. That is the reason to move it behind a gateway
-      # like the others.
-      grafana:
-        type: streamable-http
-        url: http://mcp-grafana.observability.svc.cluster.local:8000/mcp
       catalog-mcp:
         type: streamable-http
         url: https://catalog-dev.sttlab.local/catalog-mcp