Cloud deploys say 'transport error' where the CLI says a server did not answer#3927

What happens

When a deploy run by Ployz Cloud can't reach a server to remove a copy, the left-running warning gives transport error as the reason:

my-app-production/redis copy 974c89d3442f left running on MY-SERVER (transport error). The next deploy cleans it up.

The same outage run from a CLI deploy reads (MY-SERVER did not answer). The did not answer wording is chosen only when the error code is Unavailable; on the Cloud path the worker logged code internal for the outage, so the warning falls back to the generic text. (Cause inferred from the worker log, not traced.)

Expected

A server that doesn't answer reads MY-SERVER did not answer on both paths, as docs/user/troubleshooting/deployments.md shows for the CLI.

Repro

Found while proving the #2218 fix (PR #3925) on a disposable 1-Server Cluster: a copy's removal met a server whose traffic was dropped for 81 s during a Cloud deploy.

F: 3 S: 2 F×S: 6