Document ExtendedResourceToleration functionality (flyteorg#924)

* Document ExtendedResourceToleration functionality

ExtendedResourceToleration can be used to automatically apply gpu tolerations, otherwise configure Flyte.

Signed-off-by: Andrew Dye <andrewwdye@gmail.com>

* Formatting

Signed-off-by: Andrew Dye <andrewwdye@gmail.com>

* Fix punctuation

Signed-off-by: Andrew Dye <andrewwdye@gmail.com>

Signed-off-by: Andrew Dye <andrewwdye@gmail.com>

cookbook/deployment/configure_use_gpus.py

@@ -11,8 +11,9 @@
 treat machines with GPUs and machines with CPUs equally. You may want to reserve machines with GPUs for tasks
 that explicitly request GPUs. To achieve this, Flyte uses the Kubernetes concept of `taints and tolerations <https://kubernetes.io/docs/concepts/scheduling-eviction/taint-and-toleration/>`__.
 
-You can configure Flyte backend to automatically schedule your task onto a node with GPUs by tolerating specific taints.
-This configuration is controlled under generic k8s plugin configuration as can be found `here <https://github.com/flyteorg/flyteplugins/blob/5a00b19d88b93f9636410a41f81a73356a711482/go/tasks/pluginmachinery/flytek8s/config/config.go#L120>`__.
+Kubernetes can automatically apply tolerations for extended resources like GPUs using the `ExtendedResourceToleration plugin <https://kubernetes.io/docs/reference/access-authn-authz/admission-controllers/#extendedresourcetoleration>`__, enabled by default in some cloud environments. Make sure the GPU nodes are tainted with a key matching the resource name, i.e., ``key: nvidia.com/gpu``.
+
+You can also configure Flyte backend to apply specific tolerations. This configuration is controlled under generic  k8s plugin configuration as can be found `here <https://github.com/flyteorg/flyteplugins/blob/5a00b19d88b93f9636410a41f81a73356a711482/go/tasks/pluginmachinery/flytek8s/config/config.go#L120>`__.
 
 The idea of this configuration is that whenever a task that can execute on Kubernetes requests for GPUs, it automatically
 adds the matching toleration for that resource (in this case, ``gpu``) to the generated PodSpec.