I am doubtful that it can only done by PQL! But you should be able to use Machine Learning capabilities and refer below link for API Documentation: Process data model API (celonis.com)

Instead of data model refresh time, I use the MAX event date in the event / activity table. I think that's more meaningful for the end users anyway, as it shows the scope of the data rather than when the data model finished running.

@Sachin Topul  Thanks for pointing out the API documentation.

@rita.kiss12 : I am currently using the MAX event date but the issue is when the data is not loaded it won't reflect and when the user filters on events the date time changes.

You can exclude the KPI / component from filtering.

Also, if the goal is to identify data model load issues, it's best to sign up for failure notifications.

If you use Custom Monitoring, you might be able to work out some more elaborate solution.