I do also remember in one of the recent AWS outages, the google cloud compute service had lower availability due to failovers hitting all at once