Torchelastic: add support for the new error file format (#57084)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/57084
The diff adds support for new error message file format:
{
"message":"test",
"timestamp": 12
}
Test Plan:
fbcode buck test mode/dev-nosan //caffe2/test/distributed/elastic/multiprocessing/errors:api_test
example job: tsm_aivanou-torchelastic_distributed_sum_77c0b147
Reviewed By: borovsky-d, wilson100hong
Differential Revision: D28042764
fbshipit-source-id: 4d21c2319654f3460d551d91cbf48568356cf4e8
pytorch
0df57401 - Torchelastic: add support for the new error file format (#57084)
