The article makes a good job of explaining the requirement but it doesn’t actually explain why expanding OTEL to support a jdbc exporter is not the best solution. Other than “it’s not straightforward”.

Tracing is not as trivial as it seems and I would be warry of rolling out a bespoke solution.

A few questions that come to mind:

• what do you mean when you say the traces should be “stateless”?
• why a SQL database? Traces typically have evolving schemas
• why can’t this be done with a straightforward Filter implementation?