Jesse Rosenberger

I'll change this to a formatted version just to represent it more cleanly, but the Query Plan itself declares two `Parallel` `Fetch`es to run this query. I'll transfer this over...

we can always reopen.

@sterpe Your reproduction looks fair. I'm not sure what's happening with @jakedetels – though it definitely sounds familiar. I reproduced @CTimmerman's issue on Meteor with `node-inspector` and got this stack-trace...

> A customer just pointed out that errors in traces are always included if the subgraph provides them in the inline-trace plugin. (And the AS3 default behavior was to include...

@kindermax Do you think that the PR #5071 will resolve this satisfactorily?

Could be a good candidate for a Tower layer!

I've converted this to a draft so we can discuss it further, but it won't be on our review backlog for now.

Idk. This PR seems out of control. I'm going to close it and let it re-open.

This is now waiting on https://github.com/open-telemetry/opentelemetry-rust/pull/965, which we're hoping comes soon. (Suggested to be this weekend!)

The fix for this is purportedly is now in https://github.com/open-telemetry/opentelemetry-rust/releases/tag/v0.19.0. I'll see about getting the integration of that upgrade prioritized on our side.