I explain our stituation. We have a pod on OpenShift with a .NET application that uses the NEST library to connect to an Elasticsearch node. What happens is that after some idle time, the calls (any call) takes 1/2 seconds, while if the calls are made sequentially they are all fast (40/50ms).
Analyzing the TCP traffic, we noticed that the pod, after establishing the TCP connection (SYN,ACK-SYN,ACK) stays idle almost 1 second before sending the CLIENT_HELLO TLS1.3 packet to Elasticsearch.
Do you know what this might be due to?