never use google cloud systems for production is a strong statement based on one random incident. CGP customer since launch 2008 (when it was App Engine) and spend around $3k/m on it (Also using AWS and Heroku). Never had a single issue but CS responses vary a lot depending on how you funnel requests.
I don't think dismissing such incidents as random incidents is a good idea.
Although crimes and thefts are rare, all developed societies invest millions into law enforcement and justice systems to measure and mitigate them.
Although device and equipment failures are rare, both private and government systems exist to address them fairly.
It seems like it's only in the tech industry that statistical insignificance of failures is assumed and then wielded as justification to avoid even minimal levels of accountability.
> It seems like it's only in the tech industry that statistical insignificance of failures is assumed and then wielded as justification to avoid even minimal levels of accountability.
Stealing this, it's far more broadly applicable than just to this sort of billing issue.
