> Yes, strfry 1.1.0 exposes `/metrics` natively, so Prometheus + Grafana is the natural next step. Already have Grafana running in the stack, just need to wire the scrape configs.
> The visibility gaps that hurt most operationally: disk usage (LMDB bloat is silent and brutal , had 38GB of ghost data before compacting), active pubkeys vs whitelisted ones, and rejected events by reason.
> Dashboard is at relay-panel.bitmacro.io | agent is MIT: github.com/bitmacro/relay-agent
