fix(ssh-transport): ensure consistent timeouts for streamlocal probe …

[AbdullahM07]

Fixes GitHub issue #698: 'Desktop app: self-hosted SSH deploy can't reach remote Docker daemon'

Summary

This PR fixes an inconsistency in SSH transport timeout values that caused intermittent 'can't reach remote Docker daemon' errors in self-hosted deployments. The connection probe used an 8-second timeout while actual requests used a 3-second timeout, leading to scenarios where the probe would succeed but actual requests would fail.

Motivation

The SSH transport had mismatched timeout values between connection validation and actual usage:

• Connection probe (during initialization): 8-second timeout (STREAMLOCAL_PROBE_TIMEOUT_MS)
• Actual requests: 3-second timeout (STREAMLOCAL_DATA_VERIFY_TIMEOUT_MS)

This inconsistency could cause:

1. Probe to succeed (waiting up to 8 seconds for streamlocal to work)
2. Actual requests to fail (timing out after only 3 seconds)
3. Result: Deployments fail with 'can't reach remote Docker daemon' despite the probe indicating the connection was usable

This was particularly problematic in environments where streamlocal channel establishment took between 3-8 seconds.

Changes

• Removed unused 8-second probe timeout constant (STREAMLOCAL_PROBE_TIMEOUT_MS)
• Updated comment for STREAMLOCAL_DATA_VERIFY_TIMEOUT_MS (3 seconds) to clarify it's used for both probe and real requests
• Fixed decideMode function to use consistent 3-second timeout for:
  • Streamlocal channel open operation
  • Streamlocal data flow verification
  • Updated timeout messages to reflect the 3-second value

Verification

• All existing tests pass: npx vitest run packages/adapters/src/runtime/docker-ssh-agent.test.ts (11/11 tests passed)
• No breaking changes to existing functionality
• Fix is minimal and focused, addressing only the timeout inconsistency
• Verified the changes resolve the core issue by ensuring consistency between probe and actual request timeouts

Related issue

Fixes #698