Fix pending aggregates restarts: Supervisor restarts aggregate, that loses command routing and remains pending

[henry-hz]

When we start a new aggregate, and kill it, supervisor immediately restart it with the last state, replaying all the events. But when sending a new Command to the same aggregate UUID, a new aggregate is created, and updated, so we end up with two aggregates on memory, one unusable (and consumed resources to be restarted) and a new fresh one, that was started by the last command request. Using :observer.start can help to track their state and see what is going on. Pls, let me know if you need more details to reproduce it.

[henry-hz]

Restart Strategies:
:permanent - the child process is always restarted
:temporary - the child process is never restarted (not even when the supervisor’s strategy is :rest_for_one or :one_for_all)
:transient - the child process is restarted only if it terminates abnormally, i.e., with an exit reason other than :normal, :shutdown or {:shutdown, term}

If we leave as transient, the process will be pending in case of any abnormal termination.

[slashdotdash]

Thanks for the pull request & fix to this issue.

The same issue applies to the process manager supervisor as it is also using the :transient restart strategy. I'll apply the same change to this supervisor too.