The production database was running out of disk space. The logs directory was 900GB. The logs were all from a cron job that ran every minute and printed 'Cron job running...' since 2018. Nobody knew what the cron job did. We deleted the logs. The cron job is still running.