Now that we have confirmed that our monitoring stack is installed and working correctly. Let's use Ansible to deploy collectd to all our Ceph nodes, so we can start monitoring it.

Switch to the ansible directory:

    cd /etc/ansible/roles
    git clone https://github.com/fiskn/Stouts.collectd

Edit your Ansible site.yml file and add the collectd role to the plays for your mon and osd nodes so that they look like the following:

Edit group_vars/all to enter the following:

Now run your site.yml playbook:

    ansible-playbook -K site.yml

The preceding command gives the following output:

You should see from the status at the end, that Ansible has deployed collectd to all your Ceph nodes, and it has configured the collectd Ceph plugin. In Grafana, you should now be able to see your Ceph nodes showing up as available metrics. The following is one of our monitor nodes:

For example, we can now create a graph showing the number objects stored in the Ceph cluster. Create a new graph in Grafana and enter the following query:

    collectd.mon1.ceph.mon.mon1.ceph_bytes.Cluster.numObject

It's advised that you spend some time browsing through the available metrics so that you are familiar with them before proceeding to the next section.