Reliability & SLA

How We Monitor Uptime

• Canary probes every 15 minutes — an automated canary exercises the full public conversion path (Cloudflare edge → API → conversion backend) around the clock, taking the same route real requests take.
• Hourly conversion-health checks — completion rates from real traffic are monitored hourly, and anomalies alert us immediately.
• Nightly end-to-end smoke test — every night an automated smoke test converts real files through the production pipeline and fails loudly on any error.
• The monitoring is itself monitored — an independent dead-man's switch alerts us out-of-band if our monitors ever go silent.

You can check live health any time on the status page, which runs real checks against the API from your browser.

Service Objectives

These are the objectives we operate and alert against — not a measured historical figure. We don't yet publish audited historical uptime, and we won't quote a number we can't source: the status page shows live health today and will grow a historical uptime chart as monitoring data accumulates.

How the Service Degrades

Under overload the API sheds load cleanly: excess requests receive a retryable HTTP 503 with a Retry-After header instead of a hung connection, so well-behaved clients ride through spikes. The website and all client-side conversions are served from Cloudflare's global edge and keep working even if the conversion backend is briefly unavailable.

Contractual SLAs

Written, contractual SLAs with service credits are available on paid API plans — Startup and above. Email support@changethisfile.com with your plan and requirements and we'll put one in place as a contract addendum. Free and Hobby plans run on the same infrastructure with the same monitoring, but without a contractual remedy.

See plans and pricing, our security practices, and the Data Processing Addendum.