According to simplified_crontab, the bibliotheca_purchase_monitor script is supposed to run every five hours. On circulation.librarysimplified.org, the timestamp for that script indicates that it last ran on March 18th, twelve days ago. (I've run it manually since then.)

For the same server, I manually ran the bin/search_index_update script last night because it hadn't run for a long time (I don't know exactly when). It's supposed to run every six minutes, but it didn't run between last night and this morning.

On circulation.openebooks.us, the bin/axis_monitor script is supposed to run every fifteen minutes. Like bibliotheca_purchase_monitor on circulation.librarysimplified.org, it last ran on March 18th.

The database_reaper script is supposed to be run once every two hours. On circulation.librarysimplified.org this script last ran on November 6th, and on circulation.openebooks.us it last ran on March 22nd.

These problems are illustrative, not comprehensive. Something's systematically wrong with the way the script runner runs the background scripts. It's not completely ignoring the need to run the scripts, since they do run occasionally without human intervention, but either the scripts aren't actually running on schedule, or they're almost always crashing before getting any work done, in a way that's not recorded.

I say "not recorded" because even if a script crashes, the crash should be logged in the "Troubleshooting" section of the admin interface, and that's not happening.

On bplsimplye.bklynlibrary.org, the scripts seem to be running regularly and everything seems to be up to date. We also haven't heard about this problem from partners. This makes me think the problem is in the way we build or deploy our Docker image.