Skip to main content

How to validate an integration in Pipefy iPaaS

  • Nov 19, 2026 12:42 PM UTC
  • 0 replies
  • 2 views
vinicius.pereira
Community Manager

👤 For teams that track processes sending data from Pipefy to other tools
🔐 Testing requires access to Pipefy iPaaS and the systems involved
🎯 Check an integration’s output and identify where to resume after a partial run

 

Finance can see a reimbursement in its spreadsheet, but the card is still in the Approved phase. The expense looks complete to the person reading the sheet and pending to the person working from the pipe. Running the whole integration again could record the expense twice.

By tracing a reimbursement from its card to Google Sheets, you can learn to recognize a complete record, choose test data that exposes errors and work out what still needs to happen when a flow stops partway through.

 

What you will learn:

 

Define what a completed record looks like

Ask the person who uses the output how they know an expense has been recorded. In this example, three pieces of evidence must point to the same reimbursement. The spreadsheet has exactly one row with the card ID and the correct amount. The card contains that row number. Finally, the card is in the Registered phase.

Each check answers a different question. The ID distinguishes two expenses submitted by the same person. The row number lets Finance trace the card back to its entry. The phase tells the team whether work is pending. For your own process, choose a reference that lets you find the same request in both tools.

Check which requests the search should pick up. If the flow looks for approved cards at 7 a.m., check the schedule’s time zone and which cards were already approved by then. A card approved at 7:15 may simply be waiting for the next run.

 

Choose test records that expose mismatches

Write down the expected result for each test card. Use recognizable test cards and inspect the destination beforehand so you can tell new rows from existing ones. Four cases are worth testing for the reimbursement process.

  • Two approved cards. Choose different IDs and amounts. Find each ID in the spreadsheet and check its amount. Two rows that both contain the first card’s ID reveal a batch mapping problem even when the row count is correct.
  • No approved cards. Check that the search returned an empty list and the sheet stayed as it was. This helps distinguish a run with nothing to process from a failure to access the pipe.
  • An unknown cost center. If a reference table supplies the accounting code, agree with Finance on who reviews an expense without a match. Verify that the expense stays available for review and does not pick up another reimbursement’s code.
  • A spreadsheet row exists but the card is still Approved. Find the row by card ID and check whether its number reached the required card field. This tests whether the team can finish the record without entering the expense again.

 

Find the right place to resume after a failure

If a card is still Approved, start with the spreadsheet. Exactly one row with the expected ID and amount confirms that the write took place. Inspect the output of the action that wrote the row and the field that should hold its number. In this example, the card could only move once that field was filled. The flow needed to write the row, update the card and then move it to Registered.

If the row is missing, inspect the action that should have created it and the spreadsheet connection. If the flow never found the card, check the phase search and whether the service account can access the pipe where the action takes place. To check service account access, see Getting Started with Pipefy Integrations. If two rows have the same ID, pause before retrying and work with Finance to identify the correct entry.

For a manual test, read the result in the editor. For a run triggered by an active flow, open the run history and find the step where the input or output first differed from what you expected. If a failure branch exists, check what it does. The person overseeing the process needs a way to find a pending card even if the flow handled the error and reached the end.

 

Reconcile the batch after the run

When the flow processes a batch, compare the eligible card IDs with the IDs written to the spreadsheet. If two claims have the same amount, the total can still look right when one claim was recorded twice and the other was missed.

Use this check on the first run after a change to the flow and whenever someone reports a discrepancy. If the search found no eligible cards, inspect a run with approved requests to verify both the spreadsheet write and the phase change.

 

Before moving on, confirm that you understand:

☐ Which ID lets you trace the same record across both systems.

☐ How to tell a completed spreadsheet write from a pending action.

☐ How to check that each eligible card reached its destination exactly once.

 

Before testing a larger batch, check the limits on run duration and concurrent executions in the Community’s Pipefy Integration Best Practices and Usage Limits Guide.