Improve error reporting for advanced troubleshooting

For instance when running make deploy-local I get this error (second line):

CRS_INSTANCE_ID is 3670791362bf0c32199b
error: either --from-file or the combination of --docker-username, --docker-password and --docker-server is required
docker-registry secret already exists
Creating TLS certificate for registry cache

However it's unclear where that is located and what command caused that.
Moreover the error doesn't seem to cause issues and everything goes forward normally.

In an ideal world we have the path of where those errors were generated, and we also have a filtered output that only shows general information of how the deploy process is going, with maybe at the end a mention that some non-fatal errors have happened.

In both fatal and non-fatal error cases, the user (or developer) can access the full log on a file, which contains everything, to see everything in context.

At the same time a flag to see the unfiltered output (like a --debug flag) would be useful too.

Start by running make deploy-local and reading the deploy-local entry point to trace which command emits the registry error. Define how fatal and non-fatal output should be distinguished, then implement filtered progress output, a full log file, and an option for unfiltered debug output; done means the deployment remains understandable while detailed context is preserved.