fuse dashboard
Serve a local page with hosts, environments, snapshots and metrics
fuse dashboard
fuse dashboard --proxy-metrics http://127.0.0.1:7080/metricsServes a page on localhost that shows the whole fleet behind the active context, refreshed every 5 seconds (2s, 15s or paused from the page):
- Tiles: host count, running environments, snapshots and their total size,
fleet cpu and ram allocation, reconcile cycles, API 5xx count, and, with
--proxy-metrics, connected guests, open streams with bytes per second, and path migrations. - Hosts: state, region, backend and arch, running environments, and allocated against capacity for vms, cpus, ram, storage and gpus.
- Environments: every environment with its state, host, size, health and egress mode, filterable by state or by id, host or task.
- Snapshots: newest first, filterable by mode or by id, vm or name.
- Metrics: every series from the orchestrator’s
/metrics, and from fuse-proxy’s when given, with a per-second rate for counters, filterable by name or label.
It’s a dev tool. It stores nothing, and it reads only what the context’s token can already read through the API. When one source fails, for example a token that can’t list hosts, the page shows the error and renders the rest.
| Flag | Default | Description |
|---|---|---|
--listen |
127.0.0.1:0 |
Local address to serve on. :0 picks a free port. |
--proxy-metrics |
fuse-proxy /metrics url to include. |
|
--no-open |
false |
Print the url without opening a browser. |
The url carries a per-run token, as fuse desktop does, so other local processes can’t read the fleet’s state from the port.
The command runs until interrupted.
fuse-proxy’s admin listener is loopback on the proxy host by default. To include its metrics from a laptop, forward the port first:
ssh -L 7080:127.0.0.1:7080 proxy-host
fuse dashboard --proxy-metrics http://127.0.0.1:7080/metrics