23. Watch pools with metrics, logs, JFR, and thread dumps
Read pool health through active count, queue depth, completed tasks, latency, logs, Java Flight Recorder, JDK Mission Control, and thread dumps. You will connect symptoms to causes before changing code.