How does the usefulness of auto-instrumentation differ for the #frontend versus the #backend?

Hazel asks: Several years ago, back when #observability and #opentelemetry was really getting started, I didn't expect auto-instrumentation to be so big and pervasive on the backend. But now it feels like rather than manually instrumenting things, people just prefer to rely on auto instrumentation, particularly in the backend. But in the frontend, you can’t do that the same way because you have so many payload concerns and could be size concerns, and you have this internet traffic concerns. You can’t send everything or even a lot and you can’t filter when you are done, you don’t have any of the tools that the backend has grown to rely on, and that the backend ecosystem has grown around. How do you reconcile that?

Featuring: Winston Hearn, Frontend Observability Expert and Hazel Weakly, Web Developer and SRE