Home / How we decide status
How we decide status
The label on a service page is the vendor's public feed, not a guess from our office.
What we read
About every five minutes a scheduled check reads the machine-readable feed the vendor publishes: Statuspage JSON, an RSS or Atom feed, Slack's status API, Microsoft's public posts, or a vendor incidents file. The page you open uses that saved result. It does not call the vendor again.
How a label is chosen
- Operational — the feed says none, ok, or available, and it does not list an open incident.
- Minor issues — the feed says minor or degraded, or it lists an open incident with minor impact.
- Partial outage — the feed says partial or major, or an open incident is marked major.
- Major outage — the feed says critical, or an open incident is a full outage.
- Maintenance — the feed says a window is in progress.
- Status unavailable — we could not read a live public feed. We do not paint that as Operational.
A day with no named incident is not proof the tool was healthy. The recent outages page and each 30-day bar only colour a day when the feed named an incident.
What we will not do
We do not average user reports, and we do not invent a green light when Okta or anyone else hides the machine-readable feed. The official status link on the service page is the page to quote to customers.
Feed list: data sources. Mistakes we have logged: corrections.