Thursday US Admin Office Hours (CS Platform):Mar 12, 2026 - 11am PT / 12pm MT / 1pm CT / 2pm ET - Pre-session Thread

Hi Scott, hope you are well. Apparently, the rule that I ran for relationship created has caused some issues where a particular relationship has been wiped off. I need to connect with you to understand what is missing in that rule so that it dosent happens again? 

Thank you, Rishabh

Hi Scott.  I’m looking to verify information related to the Gainsight Custom Object APIs to ensure we’re on the right track.  I especially want to confirm that I’m recommending the CORRECT API to my internal teams.

1. Working with APIs in Sandbox
   1. do they work?
   2. do we need to create a new ‘Gainsight API’ connection in Sandbox

2. Production environment: Gainsight Custom Object API → Read API

We need to fetch ~40k records from a custom MDA object and want to validate our understanding of the API limits and operations.

We’ll also but using the same Read API to fetch records from other custom objects, much smaller volumes (<1,000).

GS Support documentation states:

The Read API allows you to read one or more records that match a specific criteria from the Gainsight custom object. The Read API fetches a maximum of 5000 records from one API call. 

Note: If you think there are more than 5000 records that can be fetched, pass another API call with Offset as 5001 so that remaining records (maximum 5000 records again) are fetched

2. Throttling limits 

3. 

    

Low Volume only?  Are System objects eligible for use with APIs or no?

 FYI ​@jeremy_johnston  - you MIGHT be able to go API for Read on DD objects -- I was thinking I remembered that being the case, but our documentation isn’t consistent throughout that article you are reading above.

If you go down to the “READ” section of the Custom Object API article, there are 2 important mentions that I’m currently confirming, but if correct, will make your Snowflake/data team happy if they prefer API:

1st, each back of 5,000 records can be fetched as a SINGLE API call -- screenshot below -- so you would set up, say, 10 API calls with offsets to ensure that you can pull the entire DD.

2nd, one of the Sample Request Bodies in the article calls out a nuance when using a DD as the source, making it sound like READ totally works for DD’s -- again, screenshot below, and I’m confirming.

FINALLY, THOUGH -- if the team is good with picking up a CSV from an S3 bucket, that’s honestly still the lowest-effort option and best for ensuring that the current data designer is copied promptly after processing - the final data step of the Data Designer can be set to Output to S3 -- it is that easy.

I’ll let you know when I hear back!

Scott