This article matches up with most of my own findings as well. I've been using Sidekiq for almost 10 years now. I currently use it with a project that just hit o

This article matches up with most of my own findings as well. I've been using Sidekiq for almost 10 years now. I currently use it with a project that just hit over 1 billion jobs. To me, Sidekiq is the perfect example of keeping something simple. A basic Redis + Sentinel deployment with persistence enabled and multiple read replicas has allowed me to achieve this. Scaling Sidekiq workers is also incredibly easy. I have dedicated job queues for various categories, and I simply create a new Kubernetes deployment for each category. Each deployment is set to only process a specific queue. This allows me to throttle how quickly each queue gets processed, simply based on replica/pod count.