That's true between metrics and traces, but logs use a superset of the attribute so that you can include maps pkg.go.dev/go.opentelem.... The go SDK also expressly recommends not using the OTEL logger interface directly pkg.go.dev/go.opentelem.... You're supposed to use a bridge from other libraries

An exploration of custom plugins for the OTEL collector could also be fun. Are people adding dynamic sampling control plugins, tees for centralized SIEMs, extra triggers for more detailed profiling, etc

I'd also like to see how different teams deal with security in client-side tracing. I've always been reluctant to enable that since it could be abused or spammed.

I've been considering making a really simple article just to try and explain common formatting and patterns for UCUM so others don't have to spend their time trying to parse a whole standard to write a simple unit. The LHC thankfully has a test site to translate those expressions.

I'd like to hear some real life examples of how people use the whole OTEL stack together. Parts of the standard feel really disconnected and awkward. e.g. you can't use the same attributes between traces/metrics and logs (go) and the UCUM standard has pretty bad docs (and an awful name).

Will it be on youtube music eventually?

I'd feel fine if paying the monthly fee meant that the platform didn't sell or share your harvested data.