When there are pending streams waiting to be dispatched, and all sessions in the pool end up failing to connect, then we should fail the pending streams. There was a bug there where those streams would sit in the queue indefinitely, until they either timed out or another stream was enqueued (forcing a dispatch).
When there are pending streams waiting to be dispatched, and all sessions in the pool end up failing to connect, then we should fail the pending streams. There was a bug there where those streams would sit in the queue indefinitely, until they either timed out or another stream was enqueued (forcing a dispatch).
Also add some more metadata.