In my case it would unfortunately not have helped.
2 workflows with the same name were triggered. One of them was an older version, but with the same name. So I assumed that the cronjobs are stored somewhere else than just in the SQLite database.
As a quick solution I reinstalled n8n.
Based on your description I have analyzed the situation again completely.
I think it is currently likely that 2 instances of n8n were running on different ports. So it would be possible that an old state of the workflow in the cronjob was in another instance. If I have skimmed the code of the cron node correctly, the job will be executed as it was passed when it was last saved (this n8n instance) and not as it is currently in the SQLite database.
This helps for the next time.