That seems... fine? The terms basically imply that if you publish a benchmark you need to let CRDB reproduce your benchmark and discuss it publicly