Readiness gates

Go-live should be gated by evidence, not confidence alone.

Use these gates before customer pilots, investor demos, and partner deployments. Each gate includes the expected pass condition and where to find evidence inside UmamiEdge.

Open readiness report API Launch runbook Checklist

Launch gates

Evidence required before production pilot.

Authentication

Google/GitHub OAuth works and stale-session recovery redirects to /login.

Evidence: /api/auth/clear?next=/login and /login

Database

Compatibility SQL, schema, RLS, and seed complete without errors.

Evidence: Supabase SQL editor success responses and generated Database types.

Node security

Telemetry accepts only valid node keys and records last_used_at.

Evidence: /nodes/keys and /api/telemetry/ingest

Routing

Inference requests produce route decisions with explainable node selection.

Evidence: /routing and /api/router/simulate

Customer gateway

OpenAI-compatible endpoints respond and can later proxy to a real runtime.

Evidence: /gateway and /api/v1/models

SLA operations

Incidents, failover simulations, compliance evidence, and deployment runs are visible.

Evidence: /sla, /failover, /compliance/evidence, /deployments

Ownership

Readiness by owner.

Application shell

ready

Owner: Frontend

Responsive header, global navigation, landing pages, operator console, trust center, executive demo, docs hub.

Next: Connect analytics and route-level loading/error boundaries before public launch.

Supabase schema

ready

Owner: Backend

SQL-safe schema, RLS policies, deterministic seed data, helper fix scripts, and global launch-readiness tables.

Next: Generate official Supabase types after running migrations and replace the bundled typed placeholder.

Auth & tenancy

watch

Owner: Platform

Google/GitHub OAuth, magic link, organization switcher, member roles, selected organization cookie.

Next: Add invitation acceptance UI, tenant switch audit events, and stricter admin-only screens.

Node operations

watch

Owner: Infra

Node registration, node API-key hashing, secure telemetry ingestion, incident records, Linux agent starter with nvidia-smi metrics.

Next: Package the node agent as a signed release with systemd install, release check, and retry/backoff telemetry queue.

Customer APIs

watch

Owner: API

OpenAI-compatible model, chat completion, and embedding endpoints with routing stubs and customer API-key concepts.

Next: Connect a real runtime proxy to vLLM/Ollama/TGI and enforce quotas per customer key.

Commercial readiness

ready

Owner: GTM

Enterprise plans, committed-capacity model, SLA page, billing-control page, site-owner portal, global corridor narrative.

Next: Attach pilot contracts, pricing calculator, and procurement-ready security package.

Trust & compliance

watch

Owner: Security

Trust center, compliance evidence route, residency policies, SLA breach records, audit event framework.

Next: Add privacy policy, terms, subprocessors, retention matrix, and exportable evidence bundle.