[cat9]

Nice of them to give you a credit, and a good decision from a customer service standpoint, but I doubt they're at fault in any real way. It could be any number of things, many of which are completely out of their control to do more than mitigate and minimize, and thus part of working with real computing systems at scale.

Cultivate healthy paranoia that systems will fail - because eventually, they will, particularly if you run 100 of them or run them for several years or any other "you have to survive 1000 coin tosses to miss the error" combinatoric series. And always make a backup before doing system-changing events like resizing a partition or reprovisioning a VM.