Docker containers stopped unexpectedly due to OrbStack auto-update restarting VM

[Lemonawa]

Describe the bug

OrbStack's Sparkle auto-updater triggered a VM restart at ~21:57 (CST) on 2026-05-21, causing all running Docker containers to receive SIGTERM and shut down without any user action.

The containers (a docker compose stack with new-api, postgres, and redis, all with restart: always) were running normally. The last log entry was at 21:57:05, after which the entire stack was killed via signal=15. Docker event logs confirm container kill → container stop → container die for all three services simultaneously, which is consistent with a VM-level restart rather than any application error.

The Sparkle updater processes (Autoupdate, Updater.app) were found still running after the incident, and the Sparkle Launcher directory timestamp matches the time of the outage.

To Reproduce

1. Run a docker compose up -d stack in OrbStack with restart: always
2. Allow OrbStack to auto-update in the background
3. Containers are stopped without warning when the VM restarts for the update

Expected behavior

OrbStack should either:

• Not restart the VM during an auto-update without user confirmation, or
• Automatically restart containers with restart: always policy after the VM comes back up (currently they stay stopped and require manual docker compose up -d)

Diagnostic report (REQUIRED)

OrbStack info:
  Version: 2.1.1
  Commit: 5938f7b0d94c0f51823fa763326a74302524c332 (v2.1.1)

System info:
  macOS: 26.4.1 (25E253)
  CPU: arm64, 10 cores
  CPU model: Apple M4
  Model: Mac16,10
  Memory: 16 GiB

Full report: https://orbstack.dev/_admin/diag/orbstack-diagreport_2026-05-22T05-56-59.518947Z.zip

Screenshots and additional context (optional)

No response