Skip to content

Dashboard & fleet Services

What this is

The Dashboard (/) is the home page after login: a fleet health at a glance view (hosts, attention, services down, alert count, maps pulse).
Reports (/reports, nav after Catalog) is history Grafana never sees: backups, OS patches, LAN live-per-day, Docker deploys/patches, console sessions — see Reports.
Fleet Services (/services) is a separate page: an icon grid of apps monitored via Uptime Kuma bindings.

Why it exists

Operators should not need to open every host to answer:

  • How many hosts exist?
  • Which ones need OS or container work?
  • Are any watched services down or TLS-expiring?
  • Where do I jump for network topology?

The dashboard answers those from last check jobs and caches (not a full SSH scan on every page load). Fleet Services exists so “apps that matter” are one filter away from Kuma, the host, or Docker.

End-to-end: first useful dashboard

  1. Add at least one server (Add a server).
  2. Run an OS or container update check if those features are on (Updates).
  3. Open / — host count and attention table should reflect checks.
  4. Optional: connect Uptime Kuma, bind monitors → /services fills.
  5. Optional: set host DNS / network settings → Network maps panel pulse moves off zero.

Done when: you can explain each tile without guessing, and empty Services is understood as “no Kuma binds yet” rather than a bug.

Dashboard

Fleet dashboard — counts, attention table, network pulse, quick links.

Dashboard (/)

Tile / block Meaning Why it is there
Servers Host count → Servers Jump to the fleet list
Services Monitored apps (Kuma bindings) → /services Apps, not only hosts
Need attention Hosts with OS/image updates or reboot pending Work queue for the week
Reboot pending Count of hosts flagging reboot-required After kernel/OS upgrades
Open alerts Inbox count → Notifications (bell) Unread problems
OS / container summary cards Aggregate package/image update counts Fleet-wide patch pressure
Needs attention table Only hosts that need work; Open / Docker shortcuts Act without hunting
Network maps panel Constellation + named/mapped/NPM counts Homelab topology entry
Quick links Reports (history), Servers, notifications, audit, Settings, Pi-hole, Certificates, Catalog Frequent destinations
Hosts map / Path map links /dns/physical#map · /dns/logical#map Open the SVG graph (not list-only)

Status comes from last check jobs (and related caches) — not continuous SSH on every open. The Network maps panel uses a cheap pulse (counts only), not a full SVG build.

Pins & host jump

For multi-host fleets, pin frequent pages with the ★ next to feature titles / Network map cards, then open them from the header ★ menu. On a host feature page, use Jump host to stay on Docker/Backups/Services/Files while switching machines. Full guide: Pins & host jump. Host Files button appears when PIHERDER_HOST_FILES=true (Host Files).

Pulse field Source
Named hosts Fleet servers with dns_name set
Mapped names ServiceDnsRecord rows
NPM hosts Sum of NPM integration poll proxy_host_count (matches Catalog → NPM detail; not only DNS via_proxy flags — the edge hostname itself is often via_proxy=false)

Fleet Services (/services)

Fleet-wide view of Uptime Kuma service bindings (host service or Docker project/container). Uses the shared ops-hero (up / down / TLS pulse) plus filter chips and search.

Fleet Services

Fleet Services grid with status filters and deep links.
Feature Use
Filter All · Up · Down · TLS issue
Search Name, host, location
App Open service URL (if set) — primary open-app tile
Kuma Open monitor in Kuma
Host Per-server services page
Docker Host Docker page when binding is a stack
Logo… Upload, fetch favicon, or remove (operator+)

On narrow viewports, service rows stack metadata and actions vertically so controls stay tappable (no horizontal squash).

Empty grid

If you see “No fleet services yet”:

  1. Connect Uptime Kuma.
  2. Bind HTTP/HTTPS monitors as service (or use Suggest matches).
  3. Return to /services or the Dashboard Services tile.

Per-host view: Server → Services (/servers/{id}/services) — same card pattern and mobile stacking for that host only.