With the recent changes to card-scanning, the card-table instrumentation updates are coarser, thus reducing their performance impact on the scanning code itself. We want to enable this instrumentation, suitably protected by a product manageable flag, and measure its impact on vs off. If the impact (especially of the latter) is sufficiently small, we'd want to enable this feature in product builds as a manageable diagnostic (can a diagnostic flag also be manageable? 🤔