docs(observability): add batch-gateway dashboards and alerts to llm-d (#2390)

* docs(observability): add batch-gateway dashboards and alerts to llm-d

Moves the Batch Gateway's user-facing observability assets into llm-d/llm-d
per the observability-integration proposal, following the pattern set by the
WVA migration: dashboards alongside the existing seven, the PrometheusRule as
a sibling of epp-alerting-rules.yaml.

The three dashboards had no datasource template variable at all and no uid on
their datasource references, so they would have loaded but not bound. Both are
retargeted to ${DS_PROMETHEUS} to match every other dashboard here.

The PrometheusRule is rendered from the chart's Helm template using the chart's
default thresholds. Its expressions keep a namespace matcher, unlike
epp-alerting-rules.yaml, because the batch-gateway metric names are unprefixed
and generic enough to collide with unrelated workloads.

Part of llm-d/llm-d#1983.
Closes llm-d/llm-d-batch-gateway#619 (llm-d half).

Signed-off-by: Kush Zingade <kush.zingade@gmail.com>

* docs(observability): address review on batch-gateway GC dashboard and alerting doc

- llm-d-batch-gateway-gc.json: use an empty `current` for the namespace
  variable and refresh it on dashboard load (`refresh: 1`), matching the
  sibling apiserver and processor dashboards, so it no longer defaults to
  namespace="default" instead of the guide's NAMESPACE.
- alerting.md: state the batch-gateway PrometheusRule's own label
  (`app: batch-gateway-metrics`) for bring-your-own-Prometheus ruleSelectors,
  since the `app: epp-metrics` instruction above it is specific to the EPP rule.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Qnxf2u2kSBxRM1AT7kD8BD
Signed-off-by: Kush Zingade <kush.zingade@gmail.com>

---------

Signed-off-by: Kush Zingade <kush.zingade@gmail.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>