*   `tf.data`

    *   `tf.data.Dataset.zip` now supports Python-style zipping, i.e. `Dataset.zip(a, b, c)`.
    * `tf.data.Dataset.shuffle` now supports `tf.data.UNKNOWN_CARDINALITY` When doing a "full shuffle" using  `dataset = dataset.shuffle(dataset.cardinality())`. But remember, a "full shuffle" will load the full dataset into memory so that it can be shuffled, so make sure to only use this with small datasets or datasets of small objects (like filenames).

- Reverted regression that prevented `client-go` latency metrics to be reported with a template URL to avoid label cardinality. ([#111752](https://github.com/kubernetes/kubernetes/pull/111752), [@aanm](https://github.com/aanm))