ssh.ClientConfig.Timeout only covers the TCP connect. A peer that accepts
the connection but never sends an SSH banner blocked ssh.NewClientConn
forever, wedging the system's updater goroutine and leaking the socket.
Set a deadline on the conn for the handshake and clear it afterward.
- show package updates in a separate Updates tab in the tabbed layout
- sort major version changes first in the change column
- use text-sm for version numbers
The WebSocket path reused sys.data, and CBOR leaves omitted fields
untouched, so stale values carried over (e.g. Details re-saved every
poll, SystemdServicesUpdated stuck true, old map entries kept).
sys.data is now set only by the regular update for alerts. Removes the
Wi-Fi-only workaround in UnmarshalResponse.
- Info.WiFi uses json "wf" with omitempty so systems without Wi-Fi no
longer store/broadcast "wifi":null
- move osascript exec into wifi_darwin.go; drop unused commandRunner
- share strongest-connection logic between table cell and sorting
- trim agent/wifi README
Stats.WiFi is now map[string]int8 (json "wf") holding only available
RSSI
readings; SSID and unavailable signals stay in Info.WiFi. Averages are
rounded to whole dBm.
Wi-Fi is collected only on the default 60s interval; real-time requests
reuse the last snapshot to avoid spawning osascript / dumping the BSS
cache every second.
- Let GPU Power Draw span full width when it's a lone card in the grid
- Split GPU charts into summary (power/engines) and per-GPU grids so
usage and VRAM charts stay paired side by side
restartService only looked for beszel.service, but install-hub.sh installs
the unit as beszel-hub.service, so the restart was skipped on a standard
install and the hub kept running the old binary.
Try the installer's name first and keep beszel.service as a fallback for
hand written units. Same for the OpenRC branch.
The hub waited only 5s (the request manager default) for stats over
WebSocket and closed the connection on any error. On hosts where
`zpool list` stalls (seen on Proxmox, up to its 10s timeout), collection
exceeded that limit, so the hub sent a close (code 1000) and the agent
reconnected. The refresh ran every other cycle, which caused a
disconnect roughly every 2 minutes.
Hub:
- Wait up to 30s for WebSocket stats responses.
- Keep the connection open when a stats request times out; only close
it (and fall back to SSH) for other errors.
Agent:
- After the first collection, refresh `zpool list` pool stats and
`zfs list` dataset usage in the background and serve cached values
meanwhile, so a hung utility cannot delay the stats response.
Sven van GinkelandGitHub
Anand HegdeandGitHub
Vito CappelloandGitHub
