System Status
Most status pages launch with a nine-nines figure on day one. Ours shows what the probe has actually recorded so far — and says how much data that is.
What is here, what is not
Live component readings go up as soon as the probe has them. The long-run availability average waits for its window, because an average over a short window is not an average.
If you are evaluating us right now and need availability commitments, that belongs in a contract, not on a marketing page. Enterprise agreements carry an explicit SLA. Ask us and we will send the current terms.
Live component check
Loading…
What this page will show Planned
Committed method Planned
Fixing the method before we have numbers means we cannot later pick a definition that flatters us.
Probes run from four regions every 30 seconds, against real request paths — not a health endpoint that returns OK while the product is broken.
A response slower than five seconds is a failure, even if it eventually returns 200. Degraded and unavailable are the same thing to the person waiting.
Deployment can be degraded while the API is healthy. Reporting one number for the whole system hides the outage you actually care about.
Anything over 30 minutes gets a written postmortem within five business days, including what we got wrong.
Until this page is live, tell us directly at [email protected]. Include your workspace name and roughly when it started — that is usually enough for us to find it in the logs without asking you to reproduce anything.
Enterprise customers get direct escalation contacts and a named responder during business hours, which is a commitment we can make today because it does not depend on tooling we have not built.
Connect one workspace and see every skill, rule, and memory your team has in play — in about ten minutes.
No credit card. Works with what you already run.