TCP Traffic Metrics

Questions
,
1

Subject: Unable to Retrieve TCP Traffic Metrics with Kong Gateway & TCPIngress

Hello Team,

I’m working on a use case where I need to stream TCP traffic through Kong deployed as an Ingress Controller using the TCPIngress resource.

Objective:

To monitor TCP traffic flowing through Kong and gather detailed metrics, including:

    1. Number of connection attempts (hits)
    1. Success and failure counts
    1. General observability of the TCP traffic stream (e.g., traffic volume, duration, etc.)

What I’ve Tried:

  1. Deployed Kong Ingress Controller: (kubernetes-ingress-controller/docs/api-reference.md at main · Kong/kubernetes-ingress-controller · GitHub)

Configured TCPIngress for routing TCP traffic successfully.

However, I was unable to retrieve any TCP-specific metrics from the Prometheus metrics endpoint.

  1. Tried Kong Gateway with SNI-based TLS Proxying: (How to: Proxy TCP traffic over TLS by SNI | Kong Docs)

Forwarded TCP traffic over TLS using SNI-based routing.

Still, no metrics related to TCP traffic are observed from the /metrics endpoint.

  1. Explored Official Monitoring with Prometheus and Grafana Documentation: (How to: Monitor metrics with Prometheus and Grafana | Kong Docs)

Reviewed Kong’s Prometheus plugin documentation and metrics dashboard setup with Grafana.

Most of the examples and supported metrics seem HTTP-centric (requests, latency, status codes, etc.).

Could not find any reference or support for TCP-specific metrics like connection counts or byte transfer statistics.

Request:

Could you please confirm:

Whether Kong currently supports collecting metrics for TCP traffic (streaming mode)?

If so, what is the recommended approach or plugin configuration to enable this?

Are there any known limitations or alternative solutions to get observability on TCP traffic through Kong?

Any insights, best practices, or documentation references would be highly appreciated.

Thank you.