The cockpit
The cockpit is the screen you land on. It answers one question - is anything wrong right now - and it answers it across every organization and project you belong to, not one at a time.
The screenshot uses a demo organization with fixed data, so what you see here matches what this page describes. Your own numbers will differ; the layout will not.
The banner, if there is one
Section titled “The banner, if there is one”An active incident pushes a banner above everything else, naming the service and how long it has been open. It is the only element that appears unprompted - if there is no banner, nothing is confirmed broken.
An incident stays in that banner after it is acknowledged. Acknowledging stops the escalation, not the outage; see From measurement to alarm.
The ring and the counters
Section titled “The ring and the counters”The ring shows the share of services that are healthy, with the remainder split into degraded and down. The label underneath names the worst state present, which is why a ring that is mostly green can still be labelled as an outage - one service being down is a fact about your product, not an average.
Around it:
| Counter | What it counts |
|---|---|
| Down | services with a confirmed outage right now |
| Affected | services not fully healthy, including degraded ones |
| Open incidents | incidents that have not been resolved, acknowledged or not |
| Uptime 24 h / 7 days | share of passed checks over that window |
| Incidents 24 h | incidents opened in the last day, with the average time to resolve |
Uptime here is the raw share of passed checks. It is not the availability figure from an SLA report - that one applies exclusion windows and maintenance, and is floored rather than rounded. See Evidence and retention if you need the defensible number.
Live activity
Section titled “Live activity”The feed is a stream of individual check results as they arrive, with region and latency. It is the only place in the product that shows unconfirmed, single-region results.
That is deliberate, and worth understanding: what you see here is not yet an incident. A failing line in the live feed is one region’s opinion. Whether it becomes an incident depends on the quorum - see Quorum and regions. The feed is for the person who wants to watch the mechanism work, not for deciding whether to wake somebody.
The pill in the corner shows whether the stream is connected. If it is not, the rest of the cockpit is still correct - it just stops updating by itself.
Recent status changes
Section titled “Recent status changes”Confirmed transitions, newest first: which service, from which state to which, and when. This list and the 24-hour counter above it come from the same events, so they cannot disagree.
Note the vocabulary: a change reads healthy → outage or degraded → healthy. These are confirmed states, not individual check results - the same distinction that separates this panel from the live feed.
Getting started checklist
Section titled “Getting started checklist”New organizations see a short checklist. Steps that need a higher plan are marked as such and do not count against completion, so the list can reach 100 % without anybody buying anything.
- How monitoring works - the machinery behind these numbers.
- From measurement to alarm - what happens when something breaks.
- Reading a status page - what your customers see.