[5.5.x] upgrade/resume backport

[a-palchikov]

Type of change

This PR backports the improvements made to upstream after the original (i.e. 5.5.x) PR has been merged.

• Bug fix (non-breaking change which fixes an issue)
• This change has a user-facing impact

Linked tickets and other PRs

• Refs Cannot resume an inactive or non-existing upgrade operation #1471.

• Ports Dmitri/1471 upgrade resume #1495.

TODOs

• Self-review the change
• Perform manual testing
• Address review feedback

Implementation

Testing done

error for an unstarted install operation

$ ./gravity plan resume
...
Mon May  4 17:48:34 UTC	Connecting to installer
[ERROR]: failed to find installer service. Use 'gravity install' to start new installation or 'gravity join' to join the existing cluster.
	no service unit file in /home/centos/.gravity

error for resuming on an existing cluster

# ./gravity plan resume
[ERROR]: no active operation found

explicit resumption of the upgrade operation and no operation is active:

$ ./gravity upgrade --resume
[ERROR]: no active update operation found

explicit resumption of the join operation and no operation is active:

$ ./gravity join --resume
[ERROR]: no active expand operation found

clear message on expand depending on whether the cluster is degraded or another operation is in progress:

Cluster is degraded, waiting for it to become healthy

vs

Waiting for another operation to complete at 10.128.0.104

[r0mant]

I'm using this function in my gravity rollback work, can we keep it? Unless there's another way to format operation(-s) as a table.