Improve Git LFS timeout observability
Background
There are multiple Git-LFS issues where there has been a timeout or other unknown loss of service. In many cases this has occurred on self-managed instances with varying infrastructure.
A timeout could occur at multiple places within the stack, and for many different reasons. The issue today is it is not clear via the UI or in logs why a timeout has occurred.
There are 47 open LFS issues, and ones potentially related to timeouts or logging are linked to this issue.
Proposal
Identify where a Git-LFS timeout is happening and make this information available. The MVP could be to make it available in logs only.
Granted that by nature logging a timeout will be difficult, this proposal seeks to have someway to debug these problems. It could also be increasing the verbosity of LFS transactions so that every step is logged.