protoc compiler is too old for tensorflow 2.12

tensorflow 2.12 has upgrade its protoc compiler to 3.21.9 while tensorboard still keeps 3.19.6, this will make profiler data(xplane.xxx.pb) failed to load by tensorboard.
tensorboard proto:
https://github.com/tensorflow/tensorboard/blob/56b38b75dfb993d0da6a241114bb475d072b5ab9/WORKSPACE#L138
tensorflow proto
https://github.com/tensorflow/tensorflow/blob/e8473900214fb5508699ebeafd7ea5f3ef0c2fd1/tensorflow/workspace2.bzl#L463

The error messeage is as follows:

ackages/tensorboard_plugin_profile/protobuf/diagnostics_pb2.py", line 36, in <module>
    _descriptor.FieldDescriptor(
  File "/home/subratag/mc3/envs/23ww14/lib/python3.9/site-packages/google/protobuf/descriptor.py", line 561, in __new__
    _message.Message._CheckCalledFromGeneratedFile()
TypeError: Descriptors cannot not be created directly.
If this call came from a _pb2.py file, your generated code is out of date and must be regenerated with protoc >= 3.19.0.
If you cannot immediately regenerate your protos, some other possible workarounds are:
 1. Downgrade the protobuf package to 3.20.x or lower.
 2. Set PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=python (but this will use pure-Python parsing and will be much slower). 

Start with the TensorBoard WORKSPACE line and the TensorFlow workspace2.bzl line linked in the issue, then compare their protoc versions. Reproduce the failure by loading the profiler protobuf generated file shown in the traceback; done means profiler data loads without the outdated-generated-code error.