Hi all,

I need your ideas for architecture/approach on how to trigger a Machine Learning App every time a Datamodel is loaded.

More specifically if there are new rows in a specific table. The Machine Learning App itself is also adding new data (to another table) and then reloads the model again.

My initial ideas where:

creating a Trigger + Action Flow.

Trigger based on the table which is the input for my ML App. The Action FLow itself is similar to an offical Celonis Example which triggers the App via API.

But this has following limitations:

• limit of 10000 which makes it tricky.
• Furthermore this triggers for every new row. I only want to run it once. Not thousands of times for each new row.

Scheduling by time:

I could also run the Action Flow or even the ML App directly via timeschedule and time it so that it is after sufficient time after my new data is fully processed.

This however is more unreliable as the normal process of extracting source/Transformations/datamodel load might take longer and longer.

Does anyone have an idea on how to fix the issue with the trigger or a complete different approach?