Your Metacloud build includes a connection to a data source server that provides usage information through Grafana, an open-source metrics visualization and analysis tool. Grafana provides alternate views of your usage, supplementing the information available in the Metacloud Dashboard.
Accessing and Navigating Grafana
Your Grafana dashboard is available as long as the data source is online. To access the dashboard, enter its URL in your browser:
The home page displays general information, such as the total Metacloud Hypervisors (MHVs), Metacloud Control Planes (MCPs), and instances in your Availability Zone (AZ).
A vcpu stats graph displays several measurements of virtual CPU (VCPU) usage, such as the most VCPUs allocated on a single Metacloud Hypervisor (highest_alloc) and the total number used by all instances (reserved). The adjacent Flavor Usage chart lists flavors by frequency of use.
As flavors are related to resource use, it can be helpful to view these two charts together and get an idea on how flavor use affects VCPU consumption. For example. the largest_avail_alloc measurement in the the vcpu stats graph indicates the most resource-intensive flavor that can be launched per vcpu.
Tip: To view an individual value on a graph, click the label for that value. To resume displaying all values, click the same value again.
A column of hyperlinks in the upper-left area of the home page allows you to drill down for usage information on MCPs and MHVs. A column of links in the upper-right area allow you to drill down for usage information on VMs. If your cloud uses a Ceph-backed storage, a Ceph Statistics link appears below the VM links.
Using Grafana to Help with Troubleshooting
To resolve customer issues, Metacloud Support often consults Grafana, which provides some unique data views that can be helpful for troubleshooting, as in the following examples.
Viewing Network Connections
The Conntrack Entries All Table in the MHV and MCP overview pages show all network connections to and from the servers that host and support all cloud activity. A sudden, uncharacteristic drop in connections may point to a time-stamped network event that triggered a service outage for an MHV or MCP.
Tip: To view a specific MHV, MCP, or VM on an overview page, click the drop-down list in the upper-left corner of the page.
Viewing Ceph Activity
Grafana provides rich and varied performance information about an entire Ceph deployment, as well as individual object storage daemons (OSDs) and mount points. You may be able to determine that issues with Ceph are causing slow overall performance if, for example, you see unusually high cluster latency or low input/output operations per second (IOPS).
Viewing VM Activity
You might be able to trace slow network performance to VM issues. For example, if the Disk Bytes/All table on the VM Overview page shows a high level of mebibytes (MiB), VMs may be competing for shared resources on a strained MHV, which can impact performance.
Customizing the Dashboard
You can customize your Dashboard in several ways, adding or removing tables or changing look and feel. For more information, see Grafana documentation.