There are a few of those "operations" that spew ascii-floods for reporting purposes. Unix had dmesg(1) for the longest time, and people grepped to see available system features, until /proc was discovered. Common Lisp has the ROOM function which, again, prints breakdown of memory usage in non-standardized message formats across implementations. Prolog has 'statistics'.

Almost always, if no better interface is provided, people end up parsing the human-facing reporting tools. I am glad antirez is working to improve the formatting; my advice to him is to freeze the keyword names (never to be removed or changed.) That way people can write parsers that will continue to work on future versions. Also, keep the redis version the FIRST value in the results, for obvious reasons.