Fleet: every project on one screen
Reading the Fleet page: per-project health, new issues, 24-hour trends and what happens when numbers are not fresh.
On this page
Fleet is the first screen of a workspace: every project on one page, worst first, so you can see in a few seconds which app needs you.
Summary#
Four numbers at the top:
Projects needing attention — how many projects are Critical or Needs attention, out of all.
New issues · 24h — issues seen for the first time in the last 24 hours.
Events · 24h — error events received in the last 24 hours.
Node — the status of your node. See Node health.
Project health#
Each project gets one health state, from its own errors. It is separate from the node's status.
| State | When |
|---|---|
| Critical | An issue is escalating, a new fatal error appeared, or errors spike — at least 20 events in an hour and at least five times the project's usual hourly rate over the past seven days |
| Needs attention | A new error appeared |
| Healthy | Errors arrived in the last seven days, or the project is younger than seven days, with nothing above |
| Idle | No events for seven days |
| Unknown | The node is unreachable and the project's numbers are more than 10 minutes old |
Unknown is shown on purpose rather than zeros: when the node cannot be reached, Belvador does not pretend that nothing happened.
Columns#
| Column | Shows |
|---|---|
| Health | The state above |
| Project | Name and platform |
| Monitors | The worst state of the project's monitors and how many are failing, for example "2/14 failing"; "—" when it has none (that is not the same as healthy). The summary adds Monitors failing when any project has one |
| Last 24h | A small chart of events per hour |
| New | Issues first seen in the period |
| Regressed | Resolved issues that came back |
| Events | Events received |
| Users | Distinct users affected (from setUser IDs) |
| Unresolved | Open issues |
| Top issue | The issue with the most events |
| Last event | When the latest event arrived |
Some columns hide on narrow screens.
Sort, filter, pin#
Default order: pinned projects first, then worst health, then most new issues, then name.
Sort by name, new issues, events, users, unresolved or last event.
Filter by name or address, by health and by platform. Your choice stays in the page address.
Pin on Fleet (in Project settings → General) keeps a project at the top.
When numbers are not fresh#
Fleet numbers come from summaries your node sends regularly. When they are more than five minutes old, Fleet shows their age. If the node is unreachable for longer than 10 minutes, projects turn Unknown until it is back — errors sent meanwhile are held by the edge and counted once the node delivers them.
Getting started checklist#
A new workspace shows Getting started on Fleet: turn on two-factor, invite your team and set up the node. Each item links to the right page.
Spotted something wrong or missing? Tell us: hello@belvador.dev