Can anyone clarify why segment evaluation fails under moderate load in our current environment?
I am running JMeter tests simulating 500 concurrent inbound calls. The goal is to validate API throughput for real-time attribute processing. The setup uses standard Predictive Routing segments with a mix of static and dynamic attributes.
At roughly 400 concurrent sessions, the segment evaluation begins to timeout. The Architect flow logs show a 504 Gateway Timeout when the system attempts to evaluate the customer_tier dynamic attribute via a custom data action. The data action itself responds in < 200ms when tested individually via Postman. The issue appears only under concurrent load.
Environment details:
- Genesys Cloud Release: R24.12
- Region: US East (VA)
- JMeter Version: 5.6.2
- Concurrent Users: 500
The segment configuration is straightforward. I am not using complex logical operators, just simple AND conditions. The error does not occur with 200 users. It seems the segment evaluation engine hits a connection limit or queue depth issue specific to the US East region during peak concurrency.
Here is the segment config snippet:
segment_name: "High_Value_Tier"
conditions:
- attribute: "customer_tier"
operator: "equals"
value: "platinum"
- attribute: "last_call_date"
operator: "greater_than"
value: "30_days_ago"
evaluation_mode: "real_time"
I suspect this might be related to the WebSocket connection limits for real-time attribute fetching. Has anyone seen this specific threshold behavior? I need to know if this is a hard limit or a configuration error on my end.