The primary application server experiences sustained CPU usage above 90% under moderate load, causing slow responses and timeouts. Profiling suggests a runaway background job loop consuming cycles. Throttling or optimizing that job and scaling resources are necessary.