Problem
Paradox Shop's reporting module; nightly and on-demand sales reports; currently deploys and scales as part of the same fleet as checkout, shown below as one shared pool behind one load balancer. Checkout's traffic is steady and predictable; reporting's is spiky, CPU-heavy, and completely unrelated. When a big report runs, it competes for the same instances checkout needs during a traffic spike of its own. Fix the service boundary: reporting needs its own independently-scaled pool, sized for its own profile, so a report spike can never take capacity away from checkout.
Required components
Click to add a component to the canvas.
Additional components
Browse by category. Click or drag a component onto the canvas.