Status
System status, published openly.
Live checks of the gateway, dashboard and model catalogue, uptime targets, and how incidents are reported.
Component status.
Live where we can check it; marked Not checked where we cannot.
Some systems are having problems
Checked 2026-11-06 21:16 UTC
Gateway
API gateway — US-Central (primary)
OpenAI-compatible /v1 endpoint, intelligent router
Major outageModel catalogue
Published models and list prices
OperationalProvider routing & fallback
Routing strategies, fallback chains, retry policies
Not checked
Identity & Auth
Authentication
Sign-in, sign-up, SSO/SAML, session management
Not checkedTenant-isolated database
Postgres with Row-Level Security on every tenant table
Not checkedVirtual key validation
Per-key auth, RPM/TPM enforcement, budget caps
Not checked
Billing & Credits
Credit ledger
Reserve + settle, atomic balance mutations
Not checkedStripe billing & webhooks
Checkout, credit top-ups, and webhook ingest
Not checked
Dashboard & Observability
Management dashboard
Keys, teams, analytics, guardrails, settings
OperationalObservability & logging
Request logs, request trace, alerts, latency metrics
Not checked
Provider routing
Alibaba US
Qwen, DeepSeek
Not checkedOpenAI
GPT, o-series
Not checkedMicrosoft Azure
Enterprise model endpoints
Not checkedGoogle
Gemini
Not checkedAnthropic
Claude
Not checked
Major incidents (P0/P1) get a postmortem in the changelog within 7 days. See the model catalog for what is served today.
Targets and posture.
Contractual commitments and design targets.
Contractual uptime target.
Committed in the Service Level Agreement with service credits for missed minutes.
99.9%
Gateway overhead target.
Static design target for added gateway overhead. No published latency study yet.
95 ms
Incident reporting.
Major incidents (P0 / P1) get a postmortem in the changelog within 7 days: timeline, root cause, remediation.
How we monitor.
Live health checks, deploy-time log gates, and customer spend alerts.
Health probes
This page checks the gateway, dashboard and model catalogue live, at most once a minute, and every release runs a live inference probe after deploy.Release gating
Every production revision passes a log scan, health check, and post-deploy watch window before acceptance.Customer alerts
Configure per-organization budget and spend alerts delivered to webhook, email, Slack, or Teams.
Report an issue.
Open a support ticket, subscribe to incident updates, or report a security issue.
Create a support ticket
Open a support ticket if you see routing errors or latency not reflected on this board.support@nrouter.aiSubscribe to incidents
Request email notifications when an incident is opened or resolved on the status board.support@nrouter.aiSecurity disclosures
Send security-sensitive reports directly to the security team instead of filing a public issue.security@nrouter.ai
FAQs
Is this a live status feed?
Partly. The gateway, the dashboard and the model catalogue are checked live by this page, at most once a minute, and the time of the last check is shown on the board. Rows marked Not checked have no public probe yet: a problem there reaches you through support and the changelog postmortem, not this board. We chose not to embed a third-party vendor widget (Statuspage / Better Stack) because it requires a separate account and a public page that drifts from our deploys.
How do you measure uptime?
Honestly: not continuously, and not yet as a published number. This page checks the gateway, dashboard and model catalogue live, and every release is gated by a revision-scoped log scan plus a bounded post-deploy watch window. The 99.9% figure is the target we commit to in the legal SLA, not a measured result — we publish no continuous uptime history, and a Service Credit claim is assessed from our own records for the cycle. We count only customer-impacting incidents; cosmetic and self-healing transient errors do not. Any P0/P1 gets a changelog postmortem.
Where is the incident history?
Major incidents (P0 / P1) get a postmortem in the changelog within 7 days, with timeline, root cause, and remediation.