Continuous data addition to the database presents different challenges. You will have some unoptimized records to search most of the time. Depending on the rate of insertion you may be updating the metamorph indices relatively frequently as well. That needs to taken into consideration when sizing the machine. Chkind is helpful when dealing with continuous streams of data, as it removes the need to manually check and update the index. An example of an instance when you might want to override chkind would be if you had a search peak time of day. In that case you might want to force an index update right before the peak, both to ensure best search performance, as well as avoiding the index update during peak loads.