States
A terminal in Nexus Telemetry Fleet has one of 5 states on the dashboard. Each terminal also has its own page in the dashboard, which shows a more detailed state word. The dashboard state orders the fleet, worst first. The terminal’s own word says what the terminal is doing now.
Dashboard states
The dashboard state is the same word wherever the terminal appears on the dashboard. See Views and filters for the views.
| State | Meaning | Cause shown |
|---|---|---|
| Healthy | Latency and drop rate are at the terminal’s baseline | none |
| Planned outage | Down inside a maintenance window. Alerts are paused | none |
| Warning | Latency or drop rate is above the terminal’s baseline | latency above baseline or drop rate above baseline |
| Degraded | Up, but on the backup connection, or restricted by Starlink | on backup carrier or restricted by policy |
| Outage | The connection has been down for 3 minutes or more | unreachable, or stopped at the terminal when the collector reported its own shutdown |
A maintenance window is scheduled by a user, or opened by the fleet server for a collector’s update restart. See Maintenance windows. A terminal that is still down when its window ends becomes an outage immediately, with the cause maintenance window overrun. It is counted under Outage in the count strip, and its card is badged Warning.
A policy restriction is Starlink withholding or limiting service. The terminal reports the reason and the dashboard shows it, for example Service blocked in this country, Too far from its service address or Not permitted in this area.
Terminal state words
The terminal’s own page shows one of these words, or Stabilising, Recovered or Unstable during a recovery.
| Word | Meaning |
|---|---|
| Healthy | Reporting, and at its baseline |
| Restricted | Starlink is limiting service, with the reason it reports |
| On backup | The Starlink connection dropped. Traffic is on the backup connection |
| Decommissioning | The collector is being decommissioned |
| Offline | The collector is not connected, or has been silent for longer than the Stale threshold |
Freshness and the stale threshold
The terminal’s Live view shows a freshness word beside the state word. The freshness word says how recent the last reading is. It is separate from the state. A collector sends a reading every interval, 30 seconds by default.
| Word | Meaning |
|---|---|
| measuring | The first readings are still being collected |
| streaming | Readings arrive every interval |
| waiting | The last reading is late, between 1.5 and 3 intervals old |
| stale | The collector is connected and the last reading is more than 3 intervals old |
| offline | The collector is not connected |
The Stale threshold is set under Terminals on the Settings page. The default is 1 day, and the range is 1 minute to 1 year. A stale terminal reads Outage on the dashboard and Offline on its own page.
Confirmation and flapping
A state changes only when a fault or a recovery lasts. A fault must last 3 minutes to be confirmed, and a recovery 5 minutes. A 2-minute interruption is not an outage. The time shown on the card counts from when the fault was confirmed.
While a fault is being confirmed, the terminal’s state does not change. Once the fault is confirmed, Last transition records it and the terminal gets a card with the cause.
While a recovery is being confirmed, the card is muted and badged Stabilising. A fault during those 5 minutes un-mutes the card, with no new transition and no new alert. Once the recovery is confirmed, the state returns to Healthy, Last transition reads RECOVERED, and the terminal’s own page reads Recovered with a reason such as Was offline 12 min · back online.
A connection that drops and returns repeatedly is never confirmed either way. The dashboard keeps one card for it. The terminal’s own page reads Unstable with a reason such as Flapping · 4 drops, held until stable.
See Timings for every figure and its range.