I am trying to get my API workflow to run on an API call which returns a list and then save each element of that list to my DB but I think my problem is that the list to run on doesn’t match the type of thing?
Everything looks to be paired up properly but my guess is there is something more subtle I’m missing with the way the api needs to be set up in order to work
Also, I have tried adding in the “wine_recos” to the “list to run on” field but still the same:
When you do a API call I think the type changes (It’s what I see in my app) - so the type of thing is probably at the bottom of that “type of things list” called
Sorry, one more question.
I just ran this workflow and it seems to keep going. My API has 200 recommendations in it but it saved way more than that to the db. It seems to have stopped at 1062 records. I thought maybe it would go to 1400 as there are 7 fields and maybe it was running the workflow once for each field (maybe it was but then it got rate limited?). Also, I did play around with refreshing/navigating to another page, so maybe that interrupted something.
Here is my workflow
In your “Logs” area of the Editor check you are not going to 100% resources. My guess is that the Bubble runtime killed your backend workflow when it ran out of resources. Maybe.
Where it says “Interval (seconds)” in the first screenshot - put in 1 (second) so it pauses between invocations, and just see if that makes a difference. Give Bubble the chance to manage resources.
I’ve found these sorts of problems where a backend workflow just terminates for an unknown reason to be difficult to diagnose (@alexb understands!) My suggestion is to add a “debug” column on each thing you are processing and save some processing state information on each piece of workflow, for example processing time. This might give some insight into what is going on.
So, I went back in and found that it didn’t actually stop at 1062 records, it kept going to 1,400 which is exactly 7 x 200 and 200 is the number of recommendations I have and 7 is the number of attributes per recommendation. So I think the “list to run on” is running on each attribute as opposed to each recommendation. Any idea how to solve this?