Watch
0
0
Fork
You've already forked hyperhive
0
hyperhive/docs/2-handbook/hives/dashboard.md
iris 7707a257e9 hive ui: the dashboard wears the swarm UI's chrome, and its own header, tab strip and menu mounts go
The dashboard mounts @hive/shared's AppShell, the header, nav and settings menu the swarm UI has, from a new shell.js: the hive's name for the brand, the tabs agents, rebuild queue and kept state routed by the URL hash, the theme, motion and solid-background settings in this page's own browser storage, and the shared links and user menus. The three tab panes stay plain DOM, kept hidden in index.html and moved into the shell's content column by shell.js. The tabs, the hash-to-tab parse and the menu entries sit in shell-routes.js, which npm test covers (the dashboard package gets a test script). The old header, the createTabStrip tabbar, the leet tab names, the identity headline and title code, the sticky-top wrapper, the links-menu.js and user-menu.js mounts and their CSS are removed. The docs say the plain tab names and the new chrome.
2026-11-21 10:47:37 +01:00

1.8 KiB

Dashboard layout

Part of Hives. See also: Per-agent page

The hive dashboard is the per-hive UI that hive-c0re serves at /. The swarm UI holds everything swarm-wide (the agent list, agent creation, linking external forge and matrix accounts, the controller's job graph) and is where you do day-to-day work. The hive dashboard keeps shrinking: parts of it go, bit by bit.

The header looks like the swarm UI's: the hive's name, the tabs, a settings menu for the theme, a pages menu with links to the API docs and the swarm UI, and a profile menu with your authelia settings and sign-out. The URL hash (#agents, #rebuild, #kept) picks the tab.

  • agents lists the hive's containers, one row each, sorted by name, with the agent's status and the actions on it: rebuild, start, stop, destroy. A red banner at the top says two agents hash to the same port. Rename one and rebuild.
  • rebuild queue shows the rebuild queue and each build's log.
  • kept state lists agents with kept state and no container, with their size and age. PURG3 wipes one's state directory. The list also shows an agent that's part-way through a spawn, because both look the same on disk: the row says offline and the tab carries a warning.

Per-container CPU, memory and disk figures aren't shown here. They go out as metrics and appear in Grafana, and you set an agent's CPU and memory caps per agent in the swarm UI. The dashboard shows no hive warnings either; they show up as metrics and Grafana alerts (see observability).

The container row, status badges, dialogs, the selection bar, the API endpoints and the event channel are in the dashboard reference.