Optimize memory usage on images generation

With the latest atlas update I had to bump my minikube memory from 12GB to 24GB due to:
`error: failed to receive status: rpc error: code = unavailable desc = error reading from server: eof` on build.

There might be room to optimize this process so we don't have to use so much memory.

Start by locating the image-generation build entry point and reproduce the failure described in the issue with minikube configured for 12GB and 24GB. Measure memory use during the build and trace the process that causes the gRPC EOF. Done means the image-generation build completes without requiring the increased memory allocation.

Written by the indexing model from the issue text.