Dashboards & KPIs

The operations dashboard is the single in-app view of the health of your entire conversion estate. It surfaces request volume, error rate, and latency at a glance, then lets you drill from the whole estate down to a single service and operation — all in the same UI you use to manage mappings.

Everything on the dashboard reads from pre-computed aggregation tables, never from raw logs at query time, so it loads in sub-second time even across large estates. See the aggregation read-path rule.

---

KPI tiles

The top of the dashboard is a row of KPI tiles giving an at-a-glance read on the estate:

• Request volume — call count over the selected window.
• Error rate — the share of calls that resulted in an error or fault.
• Latency — end-to-end latency, with percentiles available on drill-down.
• Rejection count — payloads rejected at the 30 MB ceiling (S2R-RUN-0413).
• Backend response mismatch — count of contract-drift events and the latest mismatch reason.
• Bad-URL attempts — runtime requests that omitted the /{direction}/ segment (S2R-RUN-0405). Zero is the expected steady state for correctly wired clients.

---

Timeline charts

Timeline charts plot traffic and errors over time so you can see trends and spot incidents. Each service exposes a request/error trend with per-bucket request count, error count, error percentage, and average latency. The bucket interval is chosen to suit the selected window (finer buckets for short windows, coarser for long ones).

This data is served by GET /admin/v1/service-metric-timeline (see the Admin API).

---

Per-service rollups

From the estate-wide view you can roll down to a single service and then a single operation. Per service/operation the dashboard shows:

• traffic (request count)
• success / error rate
• latency percentiles (p50 / p95 / p99)
• converter overhead — the conversion cost on top of backend latency
• rejection count (413)
• backend response mismatch count and the latest mismatch reason
• service-over-time charts for call count and error %

The discovery-side master discovery view complements this with per-service 24-hour / 7-day / 1-month request counts across the estate, also served from aggregation tables for sub-second performance.

---

Filtering & windows

The dashboard supports multi-select filters for environment, service, and operation, with option lists loaded from GET /admin/v1/observability-filter-options. Traffic can also be narrowed by source (runtime vs a connector such as f5); default views sum across all sources.

Time windows: 15m, 1h, 24h, 7d, 30d, 60d, plus a custom range anywhere inside the retained 60-day window. See Metrics & SLO reports for the windowing details.

---

Related pages

• Observability overview
• Metrics & SLO reports
• Logs & tracing
• Admin API — observability endpoints

---

All Specaria SOAP to REST docs