I wrote a blog post a few years ago on the same topic. Honestly, the only reason I think along those lines is because the time I spent at AWS getting that idea hammered into me — if the fallback is so good, then why isn’t it the primary? If the primary isn’t good, then why can’t we fix that instead?
The idea of avoid modes and “always do the same thing” is a hard-earned piece of wisdom at Amazon and AWS. It is also why caching is seriously frowned upon; I expanded on it and other outages and system in an article on my blog: https://asim.ihsan.io/cache-or-crash/
I wrote a blog post a few years ago on the same topic. Honestly, the only reason I think along those lines is because the time I spent at AWS getting that idea hammered into me — if the fallback is so good, then why isn’t it the primary? If the primary isn’t good, then why can’t we fix that instead?
The idea of avoid modes and “always do the same thing” is a hard-earned piece of wisdom at Amazon and AWS. It is also why caching is seriously frowned upon; I expanded on it and other outages and system in an article on my blog: https://asim.ihsan.io/cache-or-crash/