What we measure, what we target, and where to find live status when things drift.
99.5% monthly uptime on the application surface (app.gozaround.co and authenticated endpoints). Marketing pages target the same SLA but lean on Cloudflare edge caching to remain visible even during origin disruption.
Planned maintenance is announced at least 48 hours in advance via the status page and email to the org_admin / super_admin of every active organization. We aim for low-traffic windows (Sundays, UTC late-night for our majority NA customer base).
Incidents are posted to the status page within 15 minutes of detection. Resolution updates every 30 minutes during an active incident. A post-mortem is published within 5 business days of any incident at "major" severity or above.
Live at status.gozaround.co (BetterStack-backed). Subscribe via email or webhook for proactive notifications.
Application availability and incidents are tracked on the status page. Inside the platform, a nightly probe checks that the outside services the program depends on are reachable: Stripe, Mandrill, Anthropic, and AWS Textract when it is configured. Results and recent history are kept per service.
The dependency probe is an unauthenticated request to each service's API host, so it tests DNS, TLS, and reachability. It does not prove our credentials are valid or that a full request would succeed, and we don't report it as though it did. Anything narrower than the claim gets said plainly here rather than dressed up.
Application endpoint 5xx responses or response time >10s for >5 consecutive minutes. Database query failures. Auth/login broken. We don't count single-feature flakiness as full-app downtime (we report it under a sub-component on the status page).