To monitor the required metrics of cluster nodes and virtual machines, we recommend setting up your own dashboard in Grafana.

Read more about dashboard settings in the official Grafana documentation.

Creating a dashboard


  1. Enter  CreateDashboard.
  2. Create a unique dashboard name: press → enter the name in the Name field → Save.

Configuring the panel


  1. To add a panel, press .
  2. Configure the panel settings: Panel Title → Edit.
  3. In the Query field, specify the data source — graphite_src.
  4. In the Series field, specify the metric name. Metric names for nodes have the format: 1.node.<node_id>.<node_metric>, metric names for virtual machines — 1.host.<vm_id_name>.<vm_metric>.

    <node_id> — cluster node id.

    <vm_name> — virtual machine id and name. E.g., 16_wizard_gold.

    <node_metric> — cluster node metric. Examples of values:

    • cpu — CPU load;
    • df.root.free — free space in the disk root section;
    • hdd.<device>.reads — disk section read IOPS. For example, hdd.sda1.reads;
    • hdd.<device>.writes — disk section write IOPS. For example, hdd.sdb1.writes;
    • la.mid — average node load;
    • mem — free RAM;
    • net.<interface>.rx — inbound traffic. For example, net.vmbr0.rx;
    • net.<interface>.tx — outbound traffic. For example, net.vmbr0.tx;
    • uptime — node uptime.

    <vm_metric> — virtual machine metric. Examples of values:

    • cpu.* — CPU load;
    • hdd — disk space used;
    • mem.unused — free RAM;
    • net.rx.bytes — inbound traffic;
    • net.tx.bytes — outbound traffic.
  5. Enter the  menu and select the data display method in the Visualization field.
  6. Enter the  menu and specify a discretionary metric name.
  7. To save settings, go to  → enter the settings commentSave.