Tables that have become stale due to a change propagation failure resulting from an out-of-code error are now automatically reloaded. A check for stale tables is performed periodically when the HeatWave Cluster is idle. Previously, identifying and reloading stale tables was a manual process. See Change Propagation.