Add configurable env vars to health-monitoring-agent DaemonSets

[nitishsaik]

Introduce a .Values.env range block in both the nvidia and non-nvidia HMA DaemonSets so custom environment variables can be injected via values, rendered before the
built-in NODE_NAME/NODE_IP fieldRef vars. Add env: [] default in values.yaml.

What's changing and why?

The health-monitoring-agent chart previously hardcoded the container env list in both DaemonSets (health-monitoring-agent and
health-monitoring-agent-non-nvidia), so there was no way to inject additional environment variables without editing the template. This change adds a {{- range .Values.env }} block to both DaemonSets, allowing custom env vars to be supplied through Helm values. The custom vars render first, followed by the existing built-in
NODE_NAME/NODE_IP fieldRef variables (and, on the non-nvidia DaemonSet, NVIDIA_VISIBLE_DEVICES/NVIDIA_DRIVER_CAPABILITIES). A default of env: [] is added to
values.yaml so the chart's behavior is unchanged unless the user explicitly sets values.

Before/After UX

Before:
Container env was fixed to NODE_NAME and NODE_IP (plus the two NVIDIA_* vars on the non-nvidia DaemonSet). Adding an env var required editing the template
directly.

After:
Users can inject env vars via values, e.g.:

​yaml env: - name: FOO value: bar - name: BAZ value: qux ​

Rendered result (both DaemonSets):

​yaml env: - name: FOO value: "bar" - name: BAZ value: "qux" - name: NODE_NAME valueFrom: fieldRef: fieldPath: spec.nodeName - name: NODE_IP valueFrom: fieldRef: fieldPath: status.hostIP ​

With no env set (the default []), output is identical to before — no leftover template artifacts.

How was this change tested?

Rendered the chart locally with Helm 4.3.0:

• helm template with default (empty) env: renders cleanly, output matches the pre-change template with no artifacts.
• helm template --set env[0].name=FOO --set env[0].value=bar --set env[1].name=BAZ --set env[1].value=qux: confirmed custom vars render first (quoted), followed by
  NODE_NAME/NODE_IP, in both DaemonSets.
• helm lint .: passed (0 charts failed).

Are unit tests added?

No. This chart has no existing unit test harness (e.g., helm-unittest); the change was verified via helm template/helm lint rendering checks instead.

Are integration tests added?

No. This is a templating-only change to a Helm chart with no runtime logic; validated through Helm rendering as above.

Reviewer Guidelines

‼️ Merge Requirements: PRs with failing integration tests cannot be merged without justification.

One of the following must be true:

• All automated PR checks pass
• Failed tests include local run results/screenshots proving they work — verified locally via helm template (empty + populated env) and helm lint (0 failures)
• Changes are documentation-only