Dataset rebuild frequency

We've noticed something unusual with the dataset, when querying the database for an IP address, the result sometimes shows an IP address listing that was "set inactiv" a few days earlier. This would suggest that the database has been edited in order to de-list certain IP addresses but that the edits have not been incorporated into the live dataset yet. There seems to be a delay in this process for some reason, in the past the edits have been transparent i.e. we have never seen an edited record before. Previously any changes to the dataset were either immediately committed to live or at least once per day but now we can see several days delay in this dataset rebuild. We have also noticed that the size of the database has increased a lot in recent months which may have something to do with it. We can't be sure that the live dataset actually reflects the recent edits, probably not.

Whitelist included

Whilst checking the dataset of l2.APEWS.org (using the dnsbl editor), we found that there are now exclude records in the dataset, lots of them. When email servers use a dnsbl they send a lookup query but if the IP address exists as an exclude record, the lookup query will return "unlisted". It looks like thousands of IP addresses from whitelists have been included in this way so the chance of errors now is greatly reduced. Certainly we've had none to report for a while as you will have noticed. Checking some of those whitelisted IP addresses and they are those of trusted senders.