Skip to main content

G-07 Dashboards

Brick G of the Enterprise AI Platform. Technology: Navin Manager (machines and services) + portal screens (agents, LLM gateway, FinOps, SLO, security); the platform no longer ships a dashboard server (ADR-0169). French version: ../../fr/bricks/G-07-dashboards.md.

Role and responsibilities​

  • Provide platform, gateway, GPU, cost, quality and security dashboards per tenant and globally.

Position and interactions​

  • Execution path: read path.
  • Machines and services: the service monitoring tab of the Navin Manager (history in the database), the host metrics and the telemetry export to ClickHouse.
  • Agents, LLM gateway, FinOps, SLO and security: the portal screens, fed by the API (/v1/telemetry/*, /v1/finops/*).
  • Logs: OpenSearch Dashboards. Traces: Jaeger and Opik. Metrics and alerts: Prometheus and Alertmanager (the alert and recording rules are still shipped).

Open source and developed parts​

  • Open source reused: Prometheus, Alertmanager, OpenSearch Dashboards, Jaeger, Opik, ClickHouse.
  • Developed in the platform: monitoring screens of the Navin Manager and observability screens of the portal, filtered by tenant.

Data model​

Control Plane tables (schema eai): eai.usage_events, eai.telemetry_spans.

Every tenant table has tenant_id uuid NOT NULL and a row level security policy on current_setting('app.tenant_id', true)::uuid. ClickHouse usage_events is the analytics replica of eai.usage_events (ReplacingMergeTree keyed by event_id).

Contract: contracts/v1/usage-event.schema.json (schema_version 1.0). Usage is always reported with contracts/v1/usage-event.schema.json, audit with contracts/v1/audit-event.schema.json.

API endpoints​

MethodPath
GET/v1/telemetry/overview
GET/v1/finops/summary

GET /v1/telemetry/overview​

Response:

{
  "window": "24h",
  "requests": 18234,
  "error_rate": 0.012,
  "p95_latency_ms": 2140,
  "tokens": {
    "input": 9120044,
    "output": 1203400
  },
  "top_models": [
    {
      "model_id": "gpt-4o-mini",
      "share": 0.61
    }
  ]
}

GET /v1/finops/summary​

Response:

{
  "group_by": "agent",
  "currency": "EUR",
  "items": [
    {
      "key": "agt_ap",
      "estimated": "41.20",
      "measured": "38.75",
      "billed": "39.02"
    }
  ]
}

SDK examples​

Python​

from enterprise_ai import EnterpriseAI

client = EnterpriseAI(base_url="<control-plane-url>", api_key="eai_live_xxx", tenant="acme", locale="en")
client.telemetry.overview(days=1)
client.finops.summary(group_by="agent")

TypeScript​

import { EnterpriseAI } from "@enterprise-ai/sdk";

const client = new EnterpriseAI({ baseUrl: "<control-plane-url>", apiKey: process.env.EAI_API_KEY, tenant: "acme", locale: "en" });
await client.telemetry.overview({ days: 1 });
await client.finops.summary({ group_by: "agent" });

Java​

EnterpriseAIClient client = EnterpriseAIClient.builder()
    .baseUrl("<control-plane-url>")
    .apiKey(System.getenv("EAI_API_KEY"))
    .tenant("acme")
    .locale(Locale.ENGLISH)
    .build();
client.telemetry().overview(Map.of("days", 1));
client.finops().summary(Map.of("group_by", "agent"));

Go​

client, err := eai.NewClient(eai.WithBaseURL("<control-plane-url>"), eai.WithAPIKey(os.Getenv("EAI_API_KEY")), eai.WithTenant("acme"), eai.WithLocale("en"))
if err != nil { log.Fatal(err) }
ctx := context.Background()
_, err = client.Telemetry.Overview(ctx, eai.WithQuery(map[string]string{"days": "1"}))
_, err = client.FinOps.Summary(ctx, eai.WithQuery(map[string]string{"group_by": "agent"}))

Configuration​

Environment variables: EAI_CLICKHOUSE_URL, EAI_OPIK_URL.

Helm values keys (deploy/helm/enterprise-ai/values.yaml): global.monitoring.serviceMonitor, global.monitoring.prometheusRules. A Grafana already run by the client can be linked as a console with EAI_GRAFANA_URL (bring your own, never deployed by the platform because of its AGPL licence, ADR-0101).

Security​

  • SSO with Keycloak, tenant users only see their organization.
  • Tenant resolved from the credential, never from the body; PostgreSQL row level security on every tenant table.
  • Every mutation and policy decision writes a hash chained audit event (G-11).
  • Errors use the standard envelope with message_i18n en and fr.

Observability​

  • Dashboard load time and query errors.
  • W3C traceparent propagated, X-Request-Id returned, spans carry eai.tenant_id, eai.project, eai.agent_id, eai.run_id.

FinOps units​

  • Cost dashboard shows estimated, measured and billed side by side.

Failure modes and resilience​

  • A data source down shows an explicit no data panel, never zeros.

Standalone adoption​

The shipped Prometheus rules can be loaded in the client's Prometheus, and a Grafana of the client can be linked as a console (EAI_GRAFANA_URL).

Acceptance tests​

  1. The observability screens of the portal and the monitoring tab of the Navin Manager load without query error; a data source down shows an explicit no data state.

Implementation roadmap