Yup. We are working to phase it out and replace it with better solutions, but it is there and working at scale (about 5 million requests a minute across the clu

Yup. We are working to phase it out and replace it with better solutions, but it is there and working at scale (about 5 million requests a minute across the cluster, looking at graphs). Killing it slowly as we move load out of it. Pretty sure we've moved out 90% of the load over the last couple of years.