video_editing_poc/docs/dashboards.md

35 lines
1.2 KiB
Markdown

# Dashboards
This document captures the dashboard artifact for operational readiness plan item `Dashboards`.
Included artifact:
- Grafana dashboard JSON: `dashboards/video-clipping-overview-grafana.json`
- VPS AI creation prompt: `docs/grafana-dashboard-creation-prompt.md`
Dashboard coverage:
- Queue backlog: pending depth, oldest pending age, processing count, DLQ count.
- API behavior: request rate and p95 latency for non-Actuator endpoints.
- Runtime health: process CPU, system CPU, and JVM heap usage.
Data source expectations:
- Prometheus scrapes `/actuator/prometheus`.
- Database-backed queue deployments expose:
- `video_clipping_queue_pending`
- `video_clipping_queue_processing`
- `video_clipping_queue_dlq`
- `video_clipping_queue_oldest_pending_age_seconds`
- Standard Spring Boot and JVM metrics come from Micrometer Actuator export.
Import guidance:
- Create or reuse a Prometheus data source in Grafana.
- Import `dashboards/video-clipping-overview-grafana.json`.
- Set the Grafana data source variable `DS_PROMETHEUS` during import.
Operational note:
- If production workers use SQS instead of the database queue, swap the queue panels to native queue metrics while keeping the same dashboard layout and thresholds.