I took a look at the logs and what hit me is that there seems to be loads of Errors along the lines of (1406, "Data too long for column 'descr_nl' at row 1") and (1265, "Data truncated for column 'systemnrbyg' at row 1"). This leads me to believe that something has either changed on the db side or possibly in how we encode what is being sent to the db.

It might be a red herring though as those errors don't show up during the creation of these reporting pages but rather during the initial data harvest.

There is additionally a pywikibot error "db_hostname" present in our user-config.py is no longer a supported(1) which could possibly mean there is some disconnect with respect to which database is being read from and written to. Should also nopt explain why entries are being added but not removed.

There is also one encoding error which unfortunately is also preventing me from easily processing the logs

This happens for the Dutch unused images as well like [https://nl.wikipedia.org/wiki/Wikipedia:Wikiproject/Erfgoed/Nederlandse_Erfgoed_Inventarisatie/Ongebruikte_foto%27s_van_gemeentelijke_monumenten this page]. This happens at around April 7 2023. On April 3 it still was OK As resolved images stay on the page, I think the harvesting from the heritage lists fails. The indication of an error at column 'descr_nl' might point to a Dutch description. Maybe one of the Dutch desciptions is too long. The template involved might be nl:template:Tabelrij gemeentelijk monument . ~~~~