This stable release fixes an issue where the destination service is throttled after overwhelming the Kubernetes API server with node topology queries. This results in the destination service failing requests and spiking in latency. By moving to a shared informer for these queries, the information is now fetched asynchronously.