Fix Frigate Prometheus metrics & rebuild Grafana dashboard #252

Merged
eblume merged 1 commit from fix/frigate-metrics-dashboard into main 2026-02-22 18:14:18 -08:00

1 commit

Author SHA1 Message Date
51875b24b3 Fix Frigate metrics scraping and rebuild Grafana dashboard
Frigate moved from minikube to ringtail but the Prometheus scrape target
still pointed at the old cluster-internal DNS. Route via Caddy on indri
(nvr.ops.eblu.me) which already proxies to ringtail.

Rebuild the Grafana dashboard for Frigate 0.17 metrics: add GPU usage,
GPU memory, device temperature, skipped FPS, and detection event rate
panels. Update stale metric names (service_info → uptime_seconds,
detection_total_fps → detection_fps).

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 18:06:14 -08:00