Administrator operations error reporting

[pferrel]

can we store error status for engine add, update, delete so that future status will report errors that are only in the logs. We are only reporting these possible errors in logs with the etcd async engine ops.

[pferrel]

The engine ops are now async with the REST calls. This is somewhat like how we handle hctl import ... and hctl train ... as "jobs". However engine ops are tied to the Administrator, not a specific engine so they may have a different lifecycle and status info. If we stored some info with other jobs, then errors that had only gone to the server logs can be stored in a form where they can be retrieved in an hctl status engines ... call.

[pferrel]

only applies to Harness 1.0+