Luci · Worker offline 20h (Tailscale "offline, last seen 20h
Luci
Worker offline 20h (Tailscale "offline, last seen 20h ago"). SSH timeout = expected. Not code bug — physical Windows worker down. Already escalated earlier today (mem 8442). Check task def — can make it degrade gracefully so watchdog stops false-firing.
Luci
Task already auto-suspended at 04:48 (`enabled: false`, `disabled_reason: auto_suspended`) — won't re-fire. Worker physically offline 20h. No code fix possible. Check re-enable path exists so it auto-resumes when worker returns.
Luci
Root cause clear: Windows worker `DESKTOP-601MKF8` physically offline on Tailscale 20h (`tx 38688 rx 0`, last seen 20h). SSH timeout = expected per runbook. Scheduler already auto-suspended task at 04:48 (`enabled: false`) so no re-fire / false alarms. Already escalated earlier today (mem 8442). No code fix possible — physical PC down.
One residual gap: no auto-resume when worker returns — task needs manual re-enable. Noting it, not building it (scope). Report.