to stick to the above point, this wasn't a minute long outage. if you care about seconds/minutes long outages, you monitor. running on aws, hetzer, ovh, or a raspberry in a shoe box makes no difference