Sam Xie

Should we use a more explainable fence for the comment? Something like `Allocate slice once to avoid allocation for the variadic argument.`

If we allow arbitrary bytes in the string, we need to support the `bytes` type. Otherwise, the arbitrary bytes value won't be accessible when SDK reads the date from OTLP....

From today's SIG meeting: - We won’t expose the `bytes` type in API before having discussions with vendors, as they might not ready to process `bytes` attributes from OTLP ---...

Created a PR for dropping non-UTF-8 string. https://github.com/open-telemetry/opentelemetry-specification/pull/3970

Based on the comments from #3970, which was trying to drop non-UTF-8 converting and mandate Unicode in the string, **dropping non-UTF-8 converting** and **mandating Unicode** in the string are considered...

I just want to provide more info here. Sorry if I didn't make this clear in the meeting, I think `sending non-UTF-8 strings as byte` is bonded with `exposing bytes...

I have added my key (also updated the company 😉)

I don't think we need to update the code based on the main branch.

> Why not use [server.address](https://github.com/open-telemetry/semantic-conventions/blob/main/docs/registry/attributes/server.md#server-address) and [server.port](https://github.com/open-telemetry/semantic-conventions/blob/main/docs/registry/attributes/server.md#server-port)? I won't object to have `server.address` and `server.port`, but our purpose is to have a unified resource attribute that can be used to...

@lmolkova > you might find this discussion interesting: https://github.com/open-telemetry/semantic-conventions/issues/2147#issuecomment-3483332119 (suggestion to capture Oracel instance name in a dedicated attribute) We currently put oracle instance name to `service.instance.id`, is that okay?...