Performance Optimization for High-Cardinality API Extractions

[PedroBi]

Hi Adobe Support Team,

I am currently experiencing significant latency and timeout issues when attempting to extract data via the Analytics Reporting API (v2.0).

The Challenge:

• Dimensions: trackingcode and date.
• Scope: 10 different RSIDs.
• Cardinality: Each RSID contains between 3,000 and 15,000 unique tracking codes per day.
• Constraints: We cannot use Data Warehouse because several of our required segments are not compatible with DWH, and we have observed minor data discrepancies between DWH and Workspace/API reporting that we need to avoid.

Current Issues: The requests are taking too long to process, frequently hitting timeout limits or triggering 429 (Too Many Requests) errors when we try to parallelize the calls.

Questions for Technical Support:

1. Are there specific latency settings or backend optimizations that can be applied to our API user/integration to handle these larger breakdowns?
2. Are there any hidden limits on the top or limit parameters for the trackingcode dimension that are causing these delays

Looking forward to your guidance on making these extractions viable.

[Jennifer_Dungan]

Hi @PedroBi ,

The Adobe Support team doesn't usually monitor the forum... while there are Adobe employees that do come to check what's happening, the majority of the support is coming from other users, just like you.

We can help support a lot of questions about "how does metric X work", or "what is the best practice for Y", or "can someone help me with my Marketing Channel Rules / Processing Rules / Classifications / etc."

Latency issues, however, are something at the server level.. we don't have access to your suite, and we can't see anything about your API calls.

You would be better to contact Adobe Client Care directly for this type of issue.

Good Luck!

[MandyGeorge]

I second what Jen said, reaching out to customer support is probably the best course of action to get info about the API limitations.

Alternatively, you mentioned you can't use data warehouse - have you considered using the data feeds? Those provide the most granular information you can get directly into a data lake, and you can choose as many or as few columns to export as needed. If you have a data lake available, that might be another option as opposed to trying to use the API.