The yes-or-no version of the homepage status. Updated every probe cycle.
Per mirror state comes from the most recent probe pass and is shown in the banner and the board. Aggregate availability counts a mirror as reachable unless it stopped answering, so a slow mirror still reads as online.
If you read “online” here and still cannot reach the platform from your end, the bottleneck is between your Tor client and the production guard pool, not at the platform itself. Refresh the New Tor Circuit menu and retry; if three retries fail in succession, file a brief note on the editorial newsroom's onion mirror so the probe configuration can be cross-checked.
That a probe completed a request to at least one mirror on its most recent pass. It does not mean every mirror is fast, that your circuit will work, or that the service will still be answering in an hour. It is a narrow claim and worth reading narrowly.
A mirror carrying high latency is still serving. Pages arrive, they just take longer. Treating degraded as down sends people hunting for alternatives when the thing in front of them would have worked if given a minute.
The bottleneck is between your Tor client and the service, not at the service. Request a new circuit for the site from the browser menu and try again. If that fails twice, move to another mirror, which forces an entirely new path on both halves of the connection.
A cold connection to a hidden service crosses about six relays with no caching anywhere. Tens of seconds is normal and up to a minute is fine when the service is busy. Abandoning at fifteen seconds produces no information and costs you the setup work again.
Component states, the 90 day uptime history and the incident log live on the status dashboard. This page covers one question in depth rather than repeating the board.