How to Find and Report Data Discrepancy

When testing Airbyte, you may encounter a situation where the data output from Airbyte differs from the data you're ingesting into your service. During this process, it's possible to come across data discrepancies, and this article aims to guide you in identifying the reasons behind these issues and taking the necessary actions to resolve them.

Airbyte catalog has over 300 connectors, each with its own unique peculiarities. The objective of this article is to help you quickly pinpoint the root of the problem. When creating a support ticket, it's essential to provide comprehensive information to assist our team in debugging the issue effectively. 

 

The first step is to analyze the high-level data for a specific metric:

  Airbyte (Service Provider)
Email Campaign Revenue $ 8500 $ 10000

We've identified a $2500 difference between Airbyte and our example data source.

The second step is to narrow down and determine the cause of the problem. While there isn't a single correct way to resolve such issues, the following strategy often proves helpful:

  • Compare based on dates: Sometimes, the problem may be isolated to a specific date range.
Date Airbyte (Service Provider)
2023-01-01 $2000 $2500
2023-01-02 $3000 $3500
2023-02-03 $3500 $4000

Notice that all the days exhibit some missing data.

  • After identifying a problematic date, filter the data to focus solely on that date and group it by common dimensions.The next step is to filter for a single date; let's use 2023-01-01, and group by Campaign to identify any campaigns experiencing issues:
Campaigns Airbyte (Service Provider)
Shoes $2000 $2000
Hats $0 $500

 

In this simple example, we've identified that the problem is caused by a specific campaign. Armed with this information, you can begin investigating why this campaign isn't returning data from the API making manual requests. Additionally, this analysis will provide significantly more information to our team when starting the debugging process compared to the high-level metric comparison at the beginning.

 

 

Was this article helpful?
1 out of 2 found this helpful

Comments

0 comments

Please sign in to leave a comment.

Articles in this section