I’m using ADF Self-hosted Integration Runtime (SHIR) in a pipeline with a copy activity that queries a DB2 database. The query filters and retrieves records. The SHIR is connected to a backend server hosting DB2, and when I run the query using ADF's debug mode, the number of records matches the number of records I get from DBVisualizer.

However, the DB2 server has been moved to the cloud, and the SHIR is still connecting through a VPN tunnel. Now, when I run the query through ADF's copy activity, I get fewer records than when I run the same query using DBVisualizer.

I’ve checked all settings, but I can’t find the reason for the missing records. The issue seems to happen when the data exceeds a certain size (in KB) — the records are dropped, but it’s not truncating columns, just the records themselves.

If I point the linked service back to the old DB (on-premises), it works fine. Is there any way I can troubleshoot this issue, or has anyone experienced a similar problem?