Last updated:
This week, I encountered another interesting customer case that initially looked like a conventional performance problem.
Multiple multi-session VDAs, sufficient resources and seemingly functional load balancing.
Yet more and more users landed on the same session host while other VDAs were barely utilized.
🔎 What stood out?
During the investigation, I noticed that some VDAs were no longer reporting their complete load indexes.
Some CPU, RAM and disk load values were missing. As a result, the Citrix Broker no longer had a reliable basis for load balancing.
The consequence: new sessions were no longer distributed evenly across the available VDAs.
🛠️ The cause
Further investigation led me to a documented Citrix bug:
CVADHELP-28688
During initialization of the Citrix Group Policy Client-Side Extension (CseEngine), determining the VDA’s FQDN may fail.
This can prevent load indexes from being reported correctly.
Citrix resolved the issue in CVAD 2507 LTSR CU2.
💡 What should you check?
If load balancing in a CVAD environment suddenly behaves strangely:
✅ Compare load indexes across all VDAs.
✅ Check that CPU, RAM and disk load values are complete.
✅ Check DNS and FQDN resolution for affected session hosts.
✅ Investigate the Citrix Policy Engine and relevant event logs.
✅ Check VDA versions and installed cumulative updates.
🎯 My takeaway
Not every Citrix performance issue is a resource problem.
Sometimes sufficient CPU, RAM and session hosts are available, but the Broker simply does not receive the information required for sensible load balancing.
That is why it is worth examining the load indexes when VDAs are unevenly utilized.
📌 Reference: Citrix CVADHELP-28688 – CVAD 2507 LTSR CU2 Fixed Issues
