1) One of the most consistent things we've learned at FTX is that third party systems will only take you so far.
Sometimes you find ones that are _so_ good you can trust them.
But usually, they're imperfect.
Conversation
Have you considered a disaster recovery failover setup between 2 platforms (ie AWS and Azure)? Or even dividing infrastructure between the two, load balancing. 50% failure instead of 100%
1
1
Replying to
would cause huge latency to load balance because communication between them takes time
disaster recovery more possible but we'd have to write lots of features twice and there would be twice the surface area for errors
Yeah definitely some trade-offs to consider. You've got your work cut out to make it stable enough for the next market crash too!
Show additional replies, including those that may contain offensive content
Show