[general health] Add retry mechanisms whilst building docker images

**Is your feature request related to a problem? Please describe.**
When building on a high latency sytem, or in CI where requests can be throttled, the docker build can fail because we're not retrying some fetches.

**Describe the solution you'd like**
We would need to add retries to the docker builds with exponential backoffs so that they build correctly. A previous attempt at doing so was botched (see #16739)

**Describe alternatives you've considered**

**Additional context**