Network reliability improvement needed

[chcosta]

When attempting to generate sbom for a binary drop from a file server, I often have to make multiple attempts before I can successfully generate an sbom. This leads to building a lot of infrastructure in a lot of places to support what is perceived as a "flaky" tool. Additionally, we have to make different design decisions when we cannot reliably generate an SBOM from a (secured) network share and must instead ensure that all sbom generation is done against local binaries. This is particularly true on larger drops.

##[debug]Starting SBOM generation workflow.
##[debug]Using the FileListBasedFileToJsonProvider provider for the files workflow.
##[debug]Enumerating all files from listFile.
##[debug]Splitting the workflow into 8 threads.
##[debug]Running the generation workflow ...
##[error]Encountered an error while generating hash for file \\cpvsbuild\drops\drop\setup\boxs\amd64ret\kor\netfx\fulllp\netfx_fulllp.mzz: The network path was not found. : '\\cpvsbuild\drops\drop1\raw\09037.00\setup\boxs\amd64ret\kor\netfx\fulllp\netfx_fulllp.mzz'
##[error]Encountered an error while generating hash for file \\cpvsbuild\drops\drop\raw\09037.00\setup\boxs\amd64ret\ptb\netfx\fulllp\netfx_fulllp.mzz: The network path was not found. : '\\cpvsbuild\drops\drop\raw\09037.00\setup\boxs\amd64ret\ptb\netfx\fulllp\netfx_fulllp.mzz'

Adding wait / retries to file access calls would improve the reliability and ease adoption of this tool.