To really understand what's going on here, you'll need to hunt down the actual executable, whether netstat or onetstat.
If netstat is aliased to onetstat in the shell - depending on which shell you're running, the "whence" or "alias" command will tell you. A simple solution might be to run the command via the shell (/bin/sh -c netstat) rather than running netstat directly.
Another possibility is that these commands are what's known as an "external link"...a way for a UNIX Service pathname to point to a conventional executable in a z/OS dataset. If this is the case, then you very well might have the netstat/onetstat in your path and otherwise correct, but you might not have the correct STEPLIB or LNKLST concatenation. When the system exec's the target (netstat/onetstat), it doesn't find the externally-linked program, and you get the "not found".
All sorts of things can go wrong here, especially when external links are involved. There can be system (APF) authorization issues, missing modules in your STEPLIB/LNKLST, not enough memory to load the program, etc etc etc. Unfortunately UNIX Services on z/OS doesn't always interpret every possible failure code, so sometimes it's necessary to go hunting. A good first start would be to catch the exception you're getting and look for the ERRNO/ERRNO2 values - they can give you a good hint.
If you have traditional z/OS facilities, your friend is the console log...SDSF's Log function or equivalent. There very well might be an x06 abend and a CSV... message on the console that would give you the clues about what to do next.