Add support for SAGEMAKER_MAX_PAYLOAD_IN_MB

[namannandan]

Description of changes: Sagemaker batch transform sets SAGEMAKER_MAX_PAYLOAD_IN_MB env var which is currently not handled and passed to the model server. Adding support for the same based on the following PR in base toolkit: https://github.com/aws/sagemaker-inference-toolkit/pull/121

Testing:

• Unit test added in PR
• Tested locally using 763104351884.dkr.ecr.us-west-2.amazonaws.com/pytorch-inference:1.13.1-cpu-py39-ubuntu20.04-sagemaker

  $ export SAGEMAKER_MAX_PAYLOAD_IN_MB=10
  $ python /usr/local/bin/dockerd-entrypoint.py serve
  $ vim logs/ts_log.log
  2023-04-08T00:53:18,889 [INFO ] main org.pytorch.serve.servingsdk.impl.PluginsManager - Initializing plugins manager...
  2023-04-08T00:53:18,889 [INFO ] main org.pytorch.serve.servingsdk.impl.PluginsManager - Initializing plugins manager...
  2023-04-08T00:53:18,979 [INFO ] main org.pytorch.serve.ModelServer -
  Torchserve version: 0.7.0
  TS Home: /opt/conda/lib/python3.9/site-packages
  Current directory: /
  Temp directory: /home/model-server/tmp
  Metrics config path: /opt/conda/lib/python3.9/site-packages/ts/configs/metrics.yaml
  Number of GPUs: 0
  Number of CPUs: 16
  Max heap size: 15824 M
  Python executable: /opt/conda/bin/python3.9
  Config file: /etc/sagemaker-ts.properties
  Inference address: http://0.0.0.0:8080
  Management address: http://0.0.0.0:8080
  Metrics address: http://127.0.0.1:8082
  Model Store: /.sagemaker/ts/models
  Initial Models: model=/opt/ml/model
  Log dir: /logs
  Metrics dir: /logs
  Netty threads: 0
  Netty client threads: 0
  Default workers per model: 16
  Blacklist Regex: N/A
  Maximum Response Size: 6553500
  Maximum Request Size: 10485760
  Limit Maximum Image Pixels: true

  Note that Maximum request size is set to 10 * 1024 * 1024 bytes.

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[sagemaker-bot]

AWS CodeBuild CI Report

• CodeBuild project: sagemaker-pytorch-inference-toolkit-pr
• Commit ID: af9a8aec1aa30da6efa7062ae89372499c11d3d3
• Result: FAILED
• Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

[lxning]

this PR breaks the PT container env convention (see notes). It will be confusing for cx. In summary, there is no acton needed in PT container.

@namannandan pls close this PR w/o merge