Status
No pre-rendered "all systems operational" badge — the widget below makes a real fetch to api.anvat.app/healthz from your network and reports what happens.
Gateway · api.anvat.app
This check runs in your browser, from your network. Auto-refreshes every 30 seconds. If it's green here, the gateway is reachable from your edge — the most honest signal we can give.
The numbers we target for paying customers. Misses get reported transparently in the incident log below.
| Metric | Target | How we measure |
|---|---|---|
| Gateway availability | 99.9% monthly | Successful 200/ok responses to /healthz from 3 monitored regions. |
| Request error rate | < 0.5% (excluding upstream) | Anvat-originated 5xx as a share of total requests. Upstream provider errors don't count against this. |
| p50 response overhead | < 25 ms vs direct upstream | Median latency added by the gateway over a direct upstream call on the same region. |
| p99 response overhead | < 80 ms vs direct upstream | 99th-percentile gateway-added latency on the same region. |
| Streaming TTFT | Within 30 ms of upstream | Time-to-first-token over SSE matches the upstream provider's TTFT within 30 ms. |
No customer-visible incidents reported yet. When something goes wrong we'll write it here with timestamps, duration, and what we changed — same week, no spin.
/healthz from 3 regions every minute.We don't have a status-page subscription endpoint yet (small team, honest about it). For now: