fix(helm): add missing hubble metrics port to retina agent daemonset

[ckaytev]

Description

Fix Prometheus metric scraping for Hubble metrics by adding the missing hubble metrics port to the Retina agent DaemonSet.

Problem

The network-observability Service (templates/hubble/metrics-service.yaml) was configured with two ports:

• hubble port (default 9965) — for Hubble metrics
• retina port (default 10093) — for Retina metrics

However, the Retina agent DaemonSet (templates/agent/daemonset.yaml) only exposed a single port (10093). The Hubble metrics server, configured via ConfigMap to listen on port 9965 (hubble-metrics-server: ":9965"), had no corresponding containerPort declared in the pod spec. This caused Prometheus scrapes to Hubble metrics to fail.

Additionally, the targetPort for the retina port in the Service used {{ .Values.retinaPort }} instead of {{ .Values.agent.container.retina.ports.containerPort }}, which could cause inconsistency if these values diverge.

Changes

• templates/agent/daemonset.yaml (Linux & Windows): Added the hubble metrics port ({{ .Values.hubble.metrics.port }}) to the container ports list
• templates/hubble/metrics-service.yaml: Fixed targetPort for the retina port to use {{ .Values.agent.container.retina.ports.containerPort }}

Port Mapping

Component	Port Name	Port Value	Purpose
DaemonSet (Linux/Windows)	retina	{{ .Values.agent.container.retina.ports.containerPort }} (default: 10093)	Retina metrics endpoint
DaemonSet (Linux/Windows)	hubble	{{ .Values.hubble.metrics.port }} (default: 9965)	Hubble metrics endpoint
Service	retina	{{ .Values.retinaPort }} → targets retina container port	Route to Retina metrics
Service	hubble	{{ .Values.hubble.metrics.port }} → targets hubble container port	Route to Hubble metrics

[ckaytev]

@microsoft-github-policy-service agree