Problems related to zone management can be really hard to debug when you don't know exactly what happened. e.g., a tool like ipadm might report some unexpected error, but it's hard to know why if you don't know what the previous state was. It'd be really handy if sled agent logged all of the shell commands it runs related to zone configuration and management (zoneadm, but also zfs, zonecfg, svcs/svcfg, networking, etc.) It'd be great to include the exit code or termination signal, stdout, and stderr. A complete log of these commands should make it easier to try to reproduce failures during zone management (e.g., #1880). I assume this wouldn't be prohibitive in terms of performance or disk space because I don't think we'd be doing these at high frequency in production.