When checking out our Nagios plugin, I noticed that although cluster functionality is incoporated in this plugin, other functionality such as shard stats seem to be missing. In the true FOSS spirit of standing on the shoulder of giants, I decided to reuse an existing Nagios plugin. Here are the configuration steps in order to include shard stats in your Nagios dashboard. The following steps are somewhat Ubuntu-specific (also applying to Debian); your mileage can vary and it's always a good idea to check distribution-specific documentation beforehand:

1. Install Nagios 3 the usual way on a server ideally separate from any cluster node (packages nagios3 and nagios-plugins) and add a password for the Nagios default user (nagiosadmin) as part of the package installation process,
2. On the cluster nodes to be monitored, install the package nagios-nrpe-server. This will take care of basic agent functionality on the nodes to be monitored. This will also look after basic server monitoring capabilities such as resource usage and other server vital signs.
3. Copy the localhost_nagios2.cfg in the /etc/nagios3/conf.d directory to redis_node1.cfg (or a similar name) in the same directory.
4. Leaving the generic_host template intact ("use generic-service" statement), replace localhost with the address of the FQDN or IP address of the cluster node in the relevant stanzas of this file.
5. Add a further stanza (easiest done by copying an existing stanza) for this cluster node and use check_redis_stats as the relevant check_command entry.
6. In your commands.cfg file (in /etc/nagios3) add another command stanza with the command line: /usr/local/bin/check_redis_stats.pl -P <port #> -H (amend with -p for password details as required if the shard on your target node requires this). This will take care of invoking a Perl script which does the actual checking on the target node.
7. Clone the following Github repo into a directory:

  https://github.com/harisekhon/nagios-plugins

In this directory either do a full build of all the plugins or simply the perl-libs (make build or make perl-libs)

8. Copy the support libraries (directory ./lib) to a location where Perl will be able to find them (hint: perl -e "print qq(@INC)" is your friend). You want to copy the complete library directory hierarchy in order to avoid any missing modules.
9. Copy the relevant Perl scripts to /usr/local/bin (you only need check_redis_stats.pl for the above configuration).
10. Start the Nagios systemd service or restart it if it's already running to reflect the changed configuration.
11. If you haven't made any mistakes, your should be able to see working Nagios installation on Nagios server IP (don't forget to add /nagios3 to the URL :-) after login with nagiosadmin/ which includes the shard stats apart from other services as configured in redis.cfg described above.

Steps 2 - 5 have to be repeated for every cluster node which is going to be monitored by this Nagios instance.

Disclaimer: This is not a polished Nagion configuration but rather "rough and ready" as I wanted to prove the concept. For production use, you would certainly hard this and confgure the standard installation to your exact needs rather than going with this generic approach. Never mind using Ansible or Chef as an orchestration tool for much more automated deployment of this infrastructure :-).