Fix network check status report issue.

[BalaBalaYi]

What changes were proposed in this pull request?

1. Use 'success' as flag to determine whether to use 'node-check failed' or 'node-check success' for reporting.
2. Use a defined error 'NodeCheckFailedError' to distinguish 'failed' and 'node-check failed' reporting.

Why are the changes needed?

1. Should return 'node-check success' if there is 1 round check which result is success.
2. Should not use 'failed exited' status to override 'node-check failed' status if the node is failed by node-check.

Does this PR introduce any user-facing change?

No.

How was this patch tested?

Training with node check failed.

[codecov]

Codecov Report

Attention: Patch coverage is 82.50000% with 7 lines in your changes missing coverage. Please review.

Project coverage is 81.41%. Comparing base (e889a69) to head (a249fa4).
Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
dlrover/python/elastic_agent/torch/training.py	72.22%	5 Missing ⚠️
...lrover/python/tests/test_elastic_training_agent.py	90.90%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1447      +/-   ##
==========================================
+ Coverage   81.21%   81.41%   +0.19%     
==========================================
  Files         238      238              
  Lines       23183    23217      +34     
==========================================
+ Hits        18829    18901      +72     
+ Misses       4354     4316      -38     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[samplise]

lgtm