What happened?

The async BatchLogProcessor implements shutdown instead of shutdown_with_timeout. This means shutdown_with_timeout is defaulted to just return Ok(()), which is the default implementation in the LogProcessor trait.

LoggerProvider routes all calls to shutdown_with_timeout, so this means shutdown does nothing and no flushing happens.

Possibly a root cause bug is that shutdown_with_timeout has a default at all? The SpanProcessor trait does not have one.

OpenTelemetry API Version (i.e version of opentelemetry crate)

opentelemetry 0.30.0

OpenTelemetry SDK Version (i.e version of opentelemetry_sdk crate)

opentelemetry-sdk 0.30.0

What Exporter(s) are you seeing the problem on?

No response

Relevant log output

Tip

_{React with 👍 to help prioritize this issue. Please use comments to provide useful context, avoiding +1 or me too, to help us triage it. Learn more here.}