Related: Second System Effect

https://en.wikipedia.org/wiki/Second-system_effect

As an aside, I'd love the Twitter engineering teams that fixed the fail whale to one day write about this as it related to fixing Twitter's stability and performance. It seems they simultaneously made major platform shifts and managed to fix things. Maybe it was done in small pieces, piece by piece and perhaps a SOA/Microservices architecture may have enabled that to feel like a rewrite without being one all at once. I'd love to hear the story. My understanding is that in addition to platform shifts, the actual engineering teams cycled out almost completely at least once, too. Really hard changes to manage through.