It is simple to calculate the expected uptime based on past data.
But a "guarantee" is totally another concept.
How do companies manage that?
It is simple to calculate the expected uptime based on past data.
But a "guarantee" is totally another concept.
How do companies manage that?
They lie through their teeth and/or have good lawyers write them service level agreements that effectively prevent customers from seeking damages when something does break?
Of course you should start out with a well-designed and resilient architecture, but occasionally things will blow up in your face when you poke them, and as you note you can’t really predict the future based on your historical track record.