Telemetry: expose hub metrics via ServiceMonitor or document why not

Checklist items [B]:

ServiceMonitor or PodMonitor exposed, or documented justification for not exposing metrics
Application logs are written to stdout/stderr in a structured format (JSON preferred)

The chart ships neither a ServiceMonitor/PodMonitor nor a justification. JupyterHub already exposes Prometheus metrics at /hub/metrics, so a small optional ServiceMonitor template (gated on a metrics.serviceMonitor.enabled value, default off) would close this for the LGTM stack.

Logging: hub/proxy/singleuser all log to stdout/stderr, but in JupyterHub's tornado text format rather than JSON. Either configure JSON logging or add a one-line documented justification that the default format is what upstream supports.

Acceptance criteria:

• ServiceMonitor template (or written justification in the README/docs)
• Log format decision documented

Part of #120.

Start by locating the Helm chart templates and values for the optional metrics.serviceMonitor.enabled setting, then verify the existing /hub/metrics endpoint. Review the README or docs for the logging configuration and document the chosen log-format rationale. Done means a ServiceMonitor or written justification exists and the log-format decision is documented.