Monitoring

This document describes the built-in monitoring and observability component that is part of Private Cloud Director. Private Cloud Director uses open source Prometheus, Alert Manager, and Grafana as the key components of it's monitoring stack behind the scenes. The PCD monitoring system is designed to collect infrastructure & application metrics using Prometheus exporters for various PCD components and services. PCD administrators can then leverage these metrics in a variety of ways:

• Metrics visibility via built-in charts in the PCD UI
• Default dashboards provided as part of the built-in Grafana instance
• Metrics that the administrators can consume to feed into their own internal monitoring platforms

Monitoring Architecture

The Private Cloud Director monitoring system leverages three primary components:

1. Prometheus: An open-source monitoring and alerting toolkit that collects and stores metrics as time-series data.
2. Alert Manager: Handles alerts sent by the Prometheus server, including deduplicating, grouping, and routing alerts to the correct receiver.
3. Grafana: A multi-platform open-source analytics and interactive visualization web application that provides charts, graphs, and alerts when connected to supported data sources.

Region Mapping

An instance of the monitoring system that includes the three components above is provisioned per region of your Private Cloud Director setup. The metrics displayed in the grafana dashboard for a given region represent aggregated data from all virtualized clusters within that region.

Log Into Grafana

You can access the built-in grafana dashboard by following these steps:

• Log into the product_name UI and click on the home screen.
• Click on the grafana link on the home screen.
• The default login credentials are:
  • If using a new deployment of Private Cloud Director June release, you can use the same login credentials that you use for your Private Cloud Director instance to log into grafana.
  • If using a deployment of Private Cloud Director that was deployed prior to the June release, or was upgraded to the June release: The default login credentials are
    • Username: adminPassword: admin.
    • Upon first login, you will be prompted to change the default password for security purposes.

Monitored Metrics

The Private Cloud Director monitoring system tracks metrics across various categories, focusing on hypervisor health and virtual machine performance.

Hypervisor Metrics

Hypervisor metrics provide insight into the health and performance of the Private Cloud Director hosts in a given region.

Note that the identifier shown in the grafana hypervisor charts corresponds to the Host ID.

Following metrics are tracked today:

Compute Metrics

• Hypervisor CPU Total: The total CPU resources allocated to the Hypervisor
• Hypervisor Memory Total: The total memory allocated to the Hypervisor
• Hypervisor CPU Usage: The actual CPU utilization of the Hypervisor, showing host resource consumption
• Hypervisor Memory Usage: The actual memory utilization of the Hypervisor, showing hosts memory consumption
• Number of Hosts: Total number of hypervisor hosts running in the region

Storage Metrics

• Disk space ext4: Total configured disk space on the root partition across all hosts in the region.
• Disk usage ext4: Total used disk space on the root partition across all hosts in the region.
• Disk read throughput: Reports read throughput for the root partition across all hosts in the region.
• Disk write throughput: Reports write throughput for the root partition across all hosts in the region.
• Disk read throughput over time (chart): Shows a chart of read throughput for the root partition per host over time.
• Disk write throughput over time (chart): Shows a chart of write throughput for the root partition per host over time.

Networking Metrics

• Network RX throughput (chart): Shows a chart of rate of inbound traffic per host.
• Network TX throughput (chart): Shows a chart of rate of outbound traffic per host.

Virtual Machine Metrics

Virtual machine metrics provide visibility into aggregate resource utilization across all virtual machines in the given region.

Note that the identifier shown in the grafana virtual machine charts corresponds to the virtual machine ID.

Following metrics are tracked today:

VM Compute Metrics

• VM CPU Total: Total configured CPU allocated to all virtual machines in the region.
• VM Memory Total: Total configured memory across all virtual machines in the region. The configured memory of a VM refers to the amount of memory defined in its configuration settings.
• VM CPU Usage: CPU utilization across all virtual machines in this region.
• VM Memory Usage (in percentage and actual value): Memory utilized across all virtual machines in the region
• VM Memory allocated: Total memory allocated across all virtual machines
• Number of VMs: Total number of virtual machines running in the region
• CPU Throttling (percent)

VM Storage Metrics

• Total Storage: Total storage across all virtual machines
• Allocated Storage: Total allocated storage across all virtual machines
• Used Storage: Total used storage across all virtual machines
• Read Throughput: Read throughput across all virtual machines
• Write Throughput: Write throughput across all virtual machines
• Read IOPS: Read IOPs across all VMs
• Write IOPS: Write IOPs across all VMs

VM Networking Metrics

• Read Latency: Read latency across all VMs
• Write Latency: Write latency across all VMs
• RX throughput: Inbound traffic throughput across all VMs
• TX throughput: Outbound traffic throughput across all VMs
• RX Packet drop:

Dashboards

The built-in grafana instance per region includes pre-configured dashboards that display the metrics listed above.

Custom Dashboard Creation

You can create custom grafana dashboards tailored to your specific monitoring needs.

To create a custom dashboard:

1. Log in to the grafana interface through the Private Cloud Director UI
2. Navigate to the Dashboard section
3. Click "New Dashboard"
4. Add panels by selecting "Add Panel."
5. Choose visualization types and configure data sources
6. Save the dashboard with a descriptive name