Status
The status page that won’t lie to you.
Most status pages are marketing that occasionally admits an outage. Ours starts from zero: uptime history accumulates from launch day, measured in public, and we will never backfill a number we didn’t measure. Until launch, everything below is a preview of the page — not a claim.
Changelog & RSS Report something
Preview build — figures locked at launch.
Components
Four things that can break, tracked separately.
Those chips are grey on purpose. A green “operational” badge on a product that hasn’t launched would be a fabricated number, and we don’t do those — not for uptime, not for anything. The day the API goes live, the monitors go live, and the history starts recording. Day one of the graph is day one of the service.
What we will publish
The numbers, once there are numbers.
Our billing model is our incentive, disclosed: failed calls bill $0.00, so every minute of downtime is our money, not yours. Publishing the meter’s health isn’t bravery — it’s showing you the incentive doing its job.
Uptime, per component
Rolling 90-day uptime percentage for each component above, measured externally, downtime counted as it happened — not as we’d prefer to remember it.
Latency, median & p95
Rolling 90-day median and 95th-percentile latency for checks and sweeps — the same statistics we publish on the benchmarks page, kept honest continuously instead of measured once and framed.
Incident history
Every incident, dated and permanent. Nothing gets deleted when it turns embarrassing — the archive is the point.
Incident policy