Understanding where in a function the function is failing is exceptionally difficult just looking at the record history widget even with Error Messaging turned on. We had a function supposedly run 5 times, the first successfully, and then subsequently fail 4 times, but I cannot see record of that.
Hi @rosschul,
We’re actively working to provide deeper granularity and clearer error visibility across all functions and connectors. Aside from the exact step where a function fails, what other info would help you debug faster? Are you looking variable states or just a more reliable log of every attempt?
Best,
Cana
Looking for a reliable log of all attempts or even if we could select branches where we would like a log taken if it goes down a specific logic path that would be extremely useful.
We are facing a race condition when creating units in a function that are sequentially serialized. Being able to see when each function step was executed just like the test would allow better debugging.
Any update on this request? Having deeper visibility into how the functions are running and if they are failing would be extremely helpful.
Thanks,
Ross
Adding to this request: please expose the following system-generated metadata within Function logic, ideally through the Function Info datasource:
-
Function name
-
Function ID
-
Executing version or snapshot ID
-
Whether the executing reference is Development or Snapshot
We are investigating duplicate Functions and apps remaining linked to older Function snapshots after imports or restores. Knowing exactly which Function and version actually executed would significantly improve troubleshooting.
With these values available, we could write execution details to a Tulip Table alongside the calling app, app version, step, and timestamp. This would let us identify unexpected duplicates or outdated references and verify that each app is running its intended Function version.
Manually maintaining “Function Name” or “Version” variables is not reliable for this purpose. Duplicates inherit those values, and revision labels can get out of sync. The metadata needs to identify the Function and version actually executing, not what the calling app expects or the latest development version.
Native execution logging would still be valuable, but exposing this metadata would also let us build targeted diagnostics and error messages without waiting for a full logging interface. This is particularly important when promoting apps through Dev → Validation → Production.