Old Business Central PO unexpectedly resent multiple times to SAP without appearing in Celigo debug

Hi all,

We are seeing some unexpected behaviour in one of our Celigo flows and I’m wondering if anyone has experienced something similar.

Issue

PO lets call it: 226 is an older PO that has already been removed from the Celigo queue.

While testing a newer PO lets call this one 227, SAP unexpectedly started receiving PO 226 again multiple times.

The SAP team confirmed these are new incoming HTTP POST requests and not retries originating from SAP.

Example

On 25-Aug-2026, SAP received POST requests to:

/sap/ZWS_PO_GBB_SRV/PoCreateHeaderSet

at approximately:

  • 13:11:22 – 2 requests
  • 13:11:23 – 2 requests
  • 13:11:28 – 2 requests
  • 13:11:29 – 2 requests

So SAP received approximately 8 requests for the old PO 226 within a few seconds, even though we were working with PO 227 at that moment.

What we see in Celigo

We enabled Celigo debug logging while reproducing the issue.

The strange part is that:

  • The unexpected PO 226 requests are not visible in the Celigo debug logs.
  • The expected PO 227 execution is visible normally.
  • PO 226 had already been removed from the queue.
  • We are not manually retrying or re-exporting PO 226.

There is also a second issue: the final flow step is sometimes not being executed at all. There is no visible error explaining why the execution stops before that step.

Expected behaviour

Only the PO currently being processed should be sent to SAP. A PO that has already been removed from the queue should not suddenly be resent, especially multiple times.

Question

Could this be related to an internal Celigo retry mechanism, a stuck request in an internal queue, or some asynchronous retry behaviour that is not shown in the normal debug/execution logs?

I would normally create a Celigo Support ticket for this, but unfortunately the Partner Portal ticket creation is still not working properly for me, so I’m posting it here for now.

Hi Nuri,

Thank you for sharing your observations.

On this situation, what I would do is check if the integration flow uses a trigger like a scheduled export or a lookup step that could be unintentionally processing old records.

Also, you can review if the export filter or delta logic is set to only pick up newly added or changed records, rather than "all" records. This is typically managed in the Export type and Last Modified or similar fields within the flow configuration.

If data is being re-sent but does not show in Celigo debug, confirm whether any pre-processing scripts or lookup steps are stopping flows before they reach the import step, which could prevent them from being logged as runs.

Additionally, check for any recent changes to your flow’s schedule or saved searches that might have reintroduced old data into the workflow.

As for the second issue, you might want to review the flow step before the final step to see if there are any criteria or filter that can prevent records from moving into the next/final step.

I hope this helps!