24-hour /health history (sampled every 30 seconds)
24 hours ago
now
What this page measures: the public
/health endpoint of the gateway,
polled from this browser every 30 seconds. Each sample is stored in this browser's
localStorage — close the tab and history resets. For a persistent
multi-region uptime feed, point a monitoring service (UptimeRobot, Pingdom,
Datadog) at
—.
What "Degraded" means: the gateway is up but the upstream LLM provider is
returning non-200 status. The gateway will fail gracefully with a sanitized 502
until the upstream recovers — and will retry with exponential backoff on transient
errors automatically.
Found an issue? Email
[email protected]
for security;
[email protected] for ops.