feat(genai): add GenAI semantic metrics UI and backend proxy

[pristley]

Summary

Add GenAI semantic metrics support: backend proxy routes for Scouter GenAI endpoints and a Svelte UI for service/trace/span GenAI metrics (timeseries, model usage, token/cost charts), plus a TypeScript API client, strict types, tests, and documentation.

Related: closes #381

---

Files changed (by component)

• Backend routes & server

  • crates/opsml_server/src/core/genai_metrics/route.rs
  • crates/opsml_server/src/core/scouter/genai/route.rs
  • crates/opsml_server/src/core/scouter/client.rs (Routes enum additions)
  • crates/opsml_server/src/core/router.rs (router mount)
  • crates/opsml_server/src/core/mod.rs

• Frontend components

  • crates/opsml_server/opsml_ui/src/lib/components/AgentServiceDashboard/GenAIMetricsNav.svelte
  • crates/opsml_server/opsml_ui/src/lib/components/AgentServiceDashboard/GenAITimeseries.svelte
  • crates/opsml_server/opsml_ui/src/lib/components/TraceDetail/GenAIMetricsTab.svelte
  • crates/opsml_server/opsml_ui/src/lib/components/card/agent/observability/GenAiChartCard.svelte
  • crates/opsml_server/opsml_ui/src/lib/components/card/agent/observability/charts.ts

• Types & API client

  • crates/opsml_server/opsml_ui/src/lib/types/genai.ts
  • crates/opsml_server/opsml_ui/src/lib/api/genai.ts
  • crates/opsml_server/opsml_ui/src/lib/api/__tests__/genai.test.ts

• Tests (frontend)

  • crates/opsml_server/opsml_ui/src/lib/components/AgentServiceDashboard/__tests__/GenAITimeseries.test.ts
  • crates/opsml_server/opsml_ui/src/lib/components/TraceDetail/__tests__/GenAIMetricsTab.test.ts

• Documentation

  • GENAI_METRICS.md

How to test

Prereqs:

• Rust toolchain (cargo) installed via https://rustup.rs
• Node + pnpm for the UI

Run backend checks:

cargo check -p opsml_server
# or via mise
mise run format
mise run lints

Run frontend tests:

cd crates/opsml_server/opsml_ui
pnpm install
pnpm test

Manual verification:

# full dev stack (recommended)
mise run dev:both
# or frontend only
cd crates/opsml_server/opsml_ui
pnpm run dev

• Open the Agent Service dashboard -> select "GenAI Metrics".
• Confirm: Timeseries charts for tokens, latency, model usage, and cost render and respond to time-range changes.
• In a Trace detail page, open the GenAI tab and confirm per-span metrics load.

API smoke tests (curl):

curl -b "<auth-cookie>" "http://localhost:8080/api/traces/<TRACE_ID>/genai/aggregate"
curl -b "<auth-cookie>" "http://localhost:8080/api/spans/<SPAN_ID>/genai/metrics"
curl -b "<auth-cookie>" "http://localhost:8080/api/services/<SERVICE_ID>/genai/timeseries"

Screenshots / GIFs

• Timeseries dashboard and Trace GenAI tab updated UI.
• (Attach screenshots or GIFs here if available.)

Breaking changes

• None — additions only.

Notes

• Backend proxies Scouter requests and inserts AuditContext for auditing.
• Frontend client includes short TTL in-memory caching to reduce repeated fetches.
• Tests mock the internal API client and Chart.js rendering where appropriate.

Closes: #381

[pristley]

@thorrester I really loved your codebase and i did my good to contribute can you have a look and give me feedback