Feature request for Conntrack metrics

Hi Retina team,
I'm from Cosmic networking team, we want to implement SNAT port auto-increase feature based on the **real-time pod-level egress connections count**, the gauge conntrack metrics that show "a pod is keeping(open&time-wait) how many egress tcp/ucp connections at 1 timepoint" is required.

I think the dimension we need is **cluster, instance(node name), source namespace, source podname, egress destination ip, destination port, protocol(tcp/udp), value**

Really appreciate if you can help us about this. Retina team have already helped us about the alert for SNAT port exhaustion issue, and if we can launch the SNAT auto-increase feature, we can address & auto-resolve the whole SNAT issue thoroughly! This has a huge impact on improving the current COSMIC user experience.

Thanks,
Yanxin