jgrusewski
1092c931f3
feat(monitoring): add actual Scaleway billing costs, fix Argo counters
...
- Add scaleway-billing CronJob: hourly scrape of Scaleway Billing API,
pushes scaleway_billing_total_eur, scaleway_billing_consumption_eur,
and scaleway_billing_product_eur metrics to pushgateway
- Update Cluster & Costs dashboard with "Actual Costs (Scaleway Billing API)"
row: Total MTD, Compute, Storage, Network, Registry stats + product
breakdown bar chart + category pie chart
- Rename existing cost section to "Projected Costs (by node count)"
- Fix Argo Workflows dashboard: use argo_workflows_gauge for
Succeeded/Failed counts instead of cumulative counters
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-03-14 02:38:42 +01:00
jgrusewski
ddf697641a
feat(monitoring): add per-service and platform dashboards, reorganize folder structure
...
- 7 application service dashboards (Services folder): API Gateway, Trading
Service, Backtesting, Data Acquisition, ML Training, Trading Agent, Web Gateway
- 9 platform dashboards (Platform folder): Prometheus, Grafana, Loki & Promtail,
Tempo, Redis, PostgreSQL, Argo Workflows, QuestDB, Cluster & Costs
- Cluster & Costs dashboard includes Scaleway node cost estimates (€/hr, €/month)
for platform, GPU (H100), and CI compile nodes
- Reorganized existing dashboards into 7 folders: CI, Infrastructure, Operations,
Platform, Services, Trading, Training (dropped "Foxhunt" prefix from titles)
- Rewrote import.sh: removed ConfigMap deployment, switched to API-only imports
with automatic folder creation and per-dashboard folder mapping
- All 24 dashboards deployed via Grafana API (Postgres-backed)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-03-14 02:30:52 +01:00
