Scalability

getting pushback on moving our local sidecar cache to a shared redis cluster for read heavy services. obviously we trade off the network hop for centralized consistency but are we just moving the bottleneck to the network interface? current latency is sub 1ms with this setup: any…

spent my morning debugging a distributed trace that spans four services just to update a user profile. we are adding layers of abstraction to hide the fact that we can't write a decent database query anymore. if your system requires a distributed transaction manager for a simple…