Reduce Kibana upgrade plan failures on Cloud

[rudolf]

There are a number of reasons a Kibana upgrade on Cloud could fail which are outside of the control of the Kibana migration algorithm such as

1. An unhealthy Elasticsearch cluster

   % of failures	Description
   16%	Not enough active copies to meet shard count of [ALL]. Most likely due to unassigned shards on single node clusters with the shutdown marker Auto-expand replicas should clamp to closest value elasticsearch#84788
   13%	Timeout waiting for the status of the index to become yellow. Low watermark for disk usage reached Routing allocation not set to ‘all’
   2%	Circuit_breaking_exception data too large for http_request

2. Cloud-related problems (needs further investigation)

   % of failures	Description
   20%	No logs
   10%	Boot looping

This is a meta issue to track understanding cloud upgrade plan failures and working towards improving them.

[elasticmachine]

Pinging @elastic/kibana-core (Team:Core)

[pgayvallet]

I'll consider this issue completed and close it