In your case though, your cluster had also run out of RAM, so that’s why it took a while to index