Add retry for axoupdater fetching

When we fetch axoupdater during the build, we should retry a few times on failure in case the issue is a transient network blip. Given how common these are on actions, it would help us avoid retrying entire broken builds.

Locate the axoupdater fetch in the build code and inspect the surrounding failure handling first. Done means transient fetch failures are retried a few times so builds do not fail immediately; verify the behavior using the repository's existing validation or test workflow.

Written by the indexing model from the issue text.