Add configurable livenessProbe support for APISIX pod

## Context

Currently, the APISIX chart only defines a `readinessProbe` (TCP on :9080) with no `livenessProbe`.

This means that if the APISIX process freezes (no crash, but unresponsive), the pod stays in `Not Ready` state indefinitely without ever being restarted by Kubernetes.

## Expected behavior

A configurable `livenessProbe` should be available in `values.yaml`, disabled by default to avoid breaking existing deployments.

Start with values.yaml and the chart template that defines the existing readinessProbe. Trace how probe values are passed into the APISIX pod, then render or lint the Helm chart to verify that livenessProbe is configurable, disabled by default, and does not change existing deployments.

Written by the indexing model from the issue text.