[Fleet] Add workaround for upgrade_status not updated to completed if client runs a FS old version

[juliaElastic]

Related to #139704

Add fix for the edge case when clients might use the new kibana version with old FS version.
To avoid leaving the agent in upgrade_status: started state, kibana could do a check that upgraded_at > upgrade_started_at and update upgrade_status: completed as a fallback.
We could do this as a scheduled task or some other place (maybe /agent_status API that is periodically called from UI).

Alternatively the agent_status API could have its queries adjusted to include the upgraded_at > upgrade_started_at condition as a fallback for healthy status.

[elasticmachine]

Pinging @elastic/fleet (Team:Fleet)

[juliaElastic]

This issue is cancelled because the original fix is reverted.