Update Ray documentation to reflect pod template updates

Signed-off-by: Jason Parraga <sovietaced@gmail.com>

examples/ray_plugin/ray_plugin/ray_example.py

@@ -1,8 +1,7 @@
 # %% [markdown]
 # # Running Ray Tasks
 #
-# The Ray task offers the capability to execute a Ray job either on a pre-existing Ray cluster
-# or by creating a new Ray cluster using the Ray operator.
+# The Ray task offers the capability to execute a Ray job on a dynamically created Ray cluster using the Ray operator.
 #
 # :::{Warning}
 # **Version Compatibility**
@@ -51,7 +50,7 @@ def f(x):
 # Here's a breakdown of the parameters:
 #
 # - `ray_start_params`: These are the [parameters](https://docs.ray.io/en/latest/ray-core/api/doc/ray.init.html)
-#   used in the Ray `init` method, encompassing the address and object-store-memory settings.
+#   used in the Ray `init` method, encompassing the object-store-memory settings.
 # - `replicas`: Specifies the desired number of replicas for the worker group. The default is 1.
 # - `group_name`: A RayCluster can host multiple worker groups, each differentiated by its name.
 # - `runtime_env`: The [runtime environment](https://docs.ray.io/en/latest/ray-core/handling-dependencies.html#runtime-environments)
@@ -85,7 +84,8 @@ def ray_task(n: int) -> typing.List[int]:
 
 # %% [markdown]
 # :::{note}
-# The `Resources` section here is utilized to specify the resources allocated to the worker nodes.
+# By default, the `Resources` section here is utilized to specify the resources allocated to the head and nodes as well
+# as the submitter pod.
 # :::
 #
 # Lastly, define a workflow to call the Ray task.
@@ -102,6 +102,27 @@ def ray_workflow(n: int) -> typing.List[int]:
 if __name__ == "__main__":
     print(ray_workflow(n=10))
 
+# %% [markdown]
+# ## Ray Head & Worker Node Customization
+#
+# By default, the Ray plugin will base much of the configuration for the Ray head and worker node pods in the Ray
+# cluster based on the Ray job submitter pod derived from the task decorator. In some cases it is useful to customize
+# the Ray head and worker node pods and the Ray plugin supports this with optional `k8s_pod` arguments for both the
+# {py:class}`~flytekitplugins.ray.HeadNodeConfig` and {py:class}`~flytekitplugins.ray.WorkerNodeConfig` classes.
+#
+# This argument allows users to pass in a completely customized pod template. The Ray plugin will use the following
+# configurations if there are defined in the pod template:
+#
+# - Container resource requests (can also be set with the `requests` helper method)
+# - Container resource limits (can also be set with the `limits` helper method)
+# - Pod runtime class name
+#
+# :::{note}
+# Containers in the pod template must match the target container name for the Ray cluster node (ie. ray-head or
+# ray worker). Otherwise, the customized configuration will not take effect.
+# :::
+# %%
+
 # %% [markdown]
 # ## Troubleshoot
 #