@thegibson completely destroyed production Kafka cluster, more senior Ops engineers managed to fix it. But the very next day I did it again