The intention is for the tag operator to have close to zero performance impact, so that it can be left in production code. Any kind of stack analysis would be far too expensive. Also, the tag operator takes a simple string so that it’s unopinionated regarding the tagging strategy.

If you find the need for more automated, stack-based tagging, you can easily write your own operator that wraps tag, passing it whatever you deem appropriate.

That said, I am working on a Chrome DevTools extension for rxjs-spy that should hopefully make tagging less necessary and should have a bunch of new features, too.