[bug] - wslhost.exe process leak when connecting to Docker Engine via WSL interop

[tvrcgo]

Prerequisites

• I have searched for existing issues that already report this problem, without success.

Dockev Version

0.4.0

Current Behavior

When Dockev connects to a Docker Engine running inside WSL2 (via the WSL interop bridge), the wslhost.exe processes spawned on the Windows side are never terminated after each polling cycle completes. They accumulate indefinitely, eventually exhausting the Hyper-V vsock buffer and breaking all WSL2 networking with error Wsl/Service/0x80072747 (WSAENOBUFS).

Expected Behavior

wslhost.exe processes should terminate after each Docker API call completes. The process count should remain stable over time.

Steps to Reproduce

1. Install Docker Engine directly in WSL2 Ubuntu (no Docker Desktop)
2. Open Dockev, connect to the WSL2 Docker Engine
3. Leave Dockev running for several hours
4. Open Task Manager or run:

(Get-Process -Name "wslhost").Count

Additional Information

Support direct TCP connection to the Docker Engine API (e.g. tcp://127.0.0.1:2375) as an engine configuration option, bypassing the WSL interop bridge entirely. This would eliminate the wslhost.exe lifecycle issue and also improve performance.

The Docker Engine already exposes a TCP endpoint configured via daemon.json:

{
  "hosts": ["unix:///var/run/docker.sock", "tcp://127.0.0.1:2375"]
}

Dockev's "Edit Engine" dialog currently only allows selecting a WSL distro. Adding a "TCP" connection type would solve this class of issues entirely.