exclude should remove error message for process

When we have a process (or multiple processes) with an error message like this:

```bash
1.2.3.4:4704 ( 58% cpu; 20% machine; 0.476 Gbps; 0% disk IO; 2.6 GB / 8.0 GB RAM )
trace_log_file_write_error
trace_log_writer_thread_unresponsive
```

It would be useful if the `exclude` or `exclude failed` would remove the error message and the process could disappear from the process list. From a forum post [1] currently it's only possible to remove the message doing a restart of the process which doesn't work if the process/host is not reachable anymore. I also have seen this behaviour when using `fdbserver --class=test` in an ephemeral (Kubernetes) environment that the test processes keep staying in the process list and the only solution is to bounce the whole cluster (`kill; kill all; status`) at least with my current knowledge of the state. A single `kill` against the process or host doesn't work if the process is not reachable anymore.

[1] https://forums.foundationdb.org/t/how-to-remove-the-last-logged-error-msg-associated-with-a-process-in-status-output/2464

Start by tracing the `exclude` and `exclude failed` commands and how unreachable processes and their error messages are represented in `status` output. Reproduce the case with a failed or unreachable process; done means exclusion removes its error message and allows it to disappear from the process list without restarting the process or cluster.

Written by the indexing model from the issue text.