Yes, you have to keep a service log table (or field on each record, i.e. ‘IsProcessed’). In a log table you keep track of the last id processed or on the field you find all records where IsProcessed is “no”. Then you can continue where you left off. There are many posts on this topic, I’ll try to find one to link.
Correct. You have to trigger a workflow “Do While” condition is true (i.e. Do a Search for DataType (Constraint IsProcessed = ‘No’) and then have a second workflow that checks the run time of this workflow (based on some log) and starts it back up if it isn’t running.
Just one more question…
In the front end, I can set a condition that triggers my API (backend workflow)…
Is there a way to do that in the backend itself? Supouse, while register:count<0 => trigger API
Yes use schedule a workflow, schedule a workflow on a list or fire off a custom event. This is an example of a table used to track workflows and run times
Do you use schedule on a list or a recursive WF that run one item at a time? (schedule WF that schedule itself on next item until all item got processed)
The question is that I have to do a followup to see if WF was concluded… that annoys me… !!!
There should be a way that a condition in database would trigger WF itself
Nowdays, If WF is interrupted (server capacity issue), I’ve got to trigger WF manually…
That’s an absurd we got used to…
I’ll talk to Bubble Support, smthg has to change in this issue :))
did you try loop schedule? running it 1 b 1
after the schedule done running or changing a thing, step 2 is the schedule itself triggered that condition only the thing is 0. its sure its fast until 10k
backend
event workflow schedule number
action step 1 is changing 0 to 1
step 2 is the schedule that will trigger step 1 . until there is no zero on the things
Hi, if System Overflows, it breaks !!!
I’ve got to manually monitor this run…, that is no good !!!
We shall all ask Bubble to change that, too obsolet procedures for a reliable system…
Don’t you agree?
This a major problem that I would think could be resolved without too much trouble. The fact that EVERY app which needs to process data reliably has to build out workflows to check on workflows completing dictates the need for a platform solution.
For now there are workarounds described in great detail on the forums.