I've seen many people struggle with troubleshooting Kubernetes, because they couldn't for the life of them find out what's wrong with their clusters. This happens a lot. as some events are not traced to any specific cluster or microservice, or are caused by some 3rd party integration, or are too short to notice through traditional monitoring tools (i.e Grafana). For instance, short outages of 1 second each time will not show up on your dash, but will trigger an alert. Good luck finding out why you still have an alert if all the services are healthy >< anyway it's a really cool feature that addresses a major painpoint for many on-call devs.
