Compute Instance Metrics

You can monitor the health, capacity, and performance of your compute instances by using metrics, alarms, and notifications.

This topic describes the metrics emitted by the metric namespace oci_computeagent (the Compute Instance Monitoring plugin on compute instances).

You can view these metrics for individual compute instances, and for all the instances in an instance pool.

Resources: Monitoring-enabled compute instances.

Overview of Metrics for an Instance and Related Resources

This section gives an overall picture of the different types of metrics available for an instance and its storage and network devices. See the following diagram and table for a summary.

This image shows the types of metrics available for an instance and related components.

Metric Namespace Resource ID Where Measured Available Metrics
oci_computeagent Instance OCID On the instance. The metrics in this namespace are aggregated across all the related resources on the instance. For example, DiskBytesRead is aggregated across all the instance's attached storage volumes, and NetworkBytesIn is aggregated across all the instance's attached VNICs.

See Available Metrics: oci_computeagent.

oci_blockstore Boot or block volume OCID By the Block Volume service. The metrics are for an individual volume (either boot volume or block volume). See Block Volume Metrics.
oci_vcn VNIC OCID By the Networking service. The metrics are for an individual VNIC.

See VNIC Metrics.

Before You Begin

  • IAM policies: To monitor resources, you must have the required type of access in a policy  written by an administrator, whether you're using the Console or the REST API with an SDK, CLI, or other tool. The policy must give you access to the monitoring services as well as the resources being monitored. If you try to perform an action and get a message that you don't have permission or are unauthorized, confirm with your administrator the type of access you have and which compartment  you should work in. For more information about user authorizations for monitoring, see IAM Policies (Monitoring).
  • Metrics exist in Monitoring: The resources that you want to monitor must emit metrics to the Monitoring service.
  • Compute instances: To emit metrics, the Compute Instance Monitoring plugin must be enabled on the instance, and plugins must be running. The instance must also have either a service gateway or a public IP address to send metrics to the Monitoring service. For more information, see Enabling Monitoring for Compute Instances.

Available Metrics: oci_computeagent

The compute instance metrics help you measure activity level and throughput of compute instances. The metrics listed in the following table are available for any monitoring-enabled compute instance. To get these metrics, enable monitoring on the instance.

The metrics in this namespace are aggregated across all the related resources on the instance. For example, DiskBytesRead is aggregated across all the instance's attached storage volumes, and NetworkBytesIn is aggregated across all the instance's attached VNICs.

For metrics emitted by the metric namespace oci_computeagent, data points are sampled every ten seconds. A batch of six of data points is emitted every minute. Therefore, for every minute granularity, the aggregate count is always six, the aggregate sum is the sum of the six data points, and the aggregate average is the average of the six data points.

You also can use the Monitoring service to create custom queries.

Each metric includes the following dimensions :

availabilityDomain
The availability domain  where the instance resides.
faultDomain
The fault domain  where the instance resides.
imageId
The OCID of the image  for the instance.
instancePoolId
The instance pool that the instance belongs to.
region
The region  where the instance resides.
resourceDisplayName
The friendly name of the instance.
resourceId
The OCID  of the instance.
shape
The shape  of the instance.
Metric Metric Display Name Unit Description Dimensions
CpuUtilization CPU Utilization percent

Activity level from CPU. Expressed as a percentage of total time.

For instance pools, the value is averaged across all instances in the pool.

availabilityDomain

faultDomain

imageId

instancePoolId

region

resourceDisplayName

resourceId

shape

DiskBytesRead1, 3 Disk Read Bytes bytes Read throughput. Expressed as bytes read per interval.
DiskBytesWritten1, 3 Disk Write Bytes bytes Write throughput. Expressed as bytes written per interval.
DiskIopsRead1, 3 Disk Read I/O operations Activity level from I/O reads. Expressed as reads per interval.
DiskIopsWritten1, 3 Disk Write I/O operations Activity level from I/O writes. Expressed as writes per interval.
LoadAverage Load Average number of processes Average system load calculated over a 1-minute period.
MemoryAllocationStalls Memory Allocation Stalls number of stalls Number of times page reclaim was called directly.
MemoryUtilization1 Memory Utilization percent

Space currently in use. Measured by pages. Expressed as a percentage of used pages.

For instance pools, the value is averaged across all instances in the pool.

NetworksBytesIn1, 2 Network Receive Bytes bytes

Network receipt throughput. Expressed as bytes received.

NetworksBytesOut1, 2 Network Transmit Bytes bytes Network transmission throughput. Expressed as bytes transmitted.

1This metric is a cumulative counter that shows monotonically increasing behavior for each session of the Oracle Cloud Agent software, resetting when the operating system is restarted.

2The Networking service provides more metrics (in the oci_vcn metric namespace) for each VNIC on the instance. For more information, see Networking Metrics.

3The Block Volume service provides more metrics (in the oci_blockstore metric namespace) for each volume attached to the instance. For more information, see Block Volume Metrics.

Available Metrics: gpu_infrastructure_health

The compute instance metrics help you measure the activity level and throughput of compute instances. The metrics listed in the following table are available for any monitoring-enabled compute instance. To get these metrics, enable monitoring on the instance.

The metrics in this namespace are aggregated across all the related resources on the instance. For example, DiskBytesRead is aggregated across all the instance's attached storage volumes, and NetworkBytesIn is aggregated across all the instance's attached VNICs.

For metrics emitted by the metric namespace gpu_infrastructure_health, data points are sampled every ten seconds. A batch of six of data points is emitted every minute. Therefore, for every minute granularity, the aggregate count is always six, the aggregate sum is the sum of the six data points, and the aggregate average is the average of the six data points.

You also can use the Monitoring service to create custom queries.

Each metric includes the following dimensions :

component
GPU or rdma_nic
timestamp
UTC time when the payload/heartbeat is emitted
version
The payload version number for compatibility
Metric Metric Display Name Unit Description Dimensions
GpuUtilization GPU utilization percent

Activity level from GPU. Expressed as a percentage of total time.

For instance pools, the value is averaged across all instances in the pool.

availabilityDomain

faultDomain

gpuId

imageId

instancePoolId

region

resourceDisplayName

resourceId

shape

GpuMemoryUtilization GPU memory utilization percent The percentage of the GPU memory resource in use.
GpuPowerDraw GPU power draw integer The amount of GPU power used.
GpuTemperature GPU temperature integer The GPU temperature reported.
GpuEccSingleBitErrors GPU single-bit errors integer The number of GPU single bit ECC errors reported.
GpuEccDoubleBitErrors GPU double-bit errors integer The number of GPU double bit ECC errors reported.

1This metric is a cumulative counter that shows monotonically increasing behavior for each session of the Oracle Cloud Agent software, resetting when the operating system is restarted.

2The Networking service provides more metrics (in the oci_vcn metric namespace) for each VNIC on the instance. For more information, see Networking Metrics.

3The Block Volume service provides more metrics (in the oci_blockstore metric namespace) for each volume attached to the instance. For more information, see Block Volume Metrics.

Available Metrics: rdma_infrastructure_health

The compute instance metrics help you measure activity level and throughput of compute instances. The metrics listed in the following table are available for any monitoring-enabled compute instance. To get these metrics, enable monitoring on the instance.

The metrics in this namespace are aggregated across all the related resources on the instance. For example, DiskBytesRead is aggregated across all the instance's attached storage volumes, and NetworkBytesIn is aggregated across all the instance's attached VNICs.

For metrics emitted by the metric namespace rdma_infrastructure_health, data points are sampled every ten seconds. A batch of six of data points is emitted every minute. Therefore, for every minute granularity, the aggregate count is always six, the aggregate sum is the sum of the six data points, and the aggregate average is the average of the six data points.

You also can use the Monitoring service to create custom queries.

Each metric includes the following dimensions :

component
GPU or rdma_nic
timestamp
UTC time when the payload/heartbeat is emitted
version
The payload version number for compatibility
Metric Metric Display Name Unit Description Dimensions
RdmaTxBytes RDMA aggregate network transmit bytes bytes The bytes transmitted on the RDMA interface.

availabilityDomain

faultDomain

imageId

instancePoolId

rdmaId

region

resourceDisplayName

resourceId

shape

RdmaRxBytes RDMA aggregate network receive bytes bytes The bytes received on the RDMA interface.
RdmaTxPackets RDMA aggregate network transmit packets integer The number of RDMA interface packets transmitted.
RdmaRxPackets RDMA aggregate network receive packets integer The number of RDMA interface packets received.

1This metric is a cumulative counter that shows monotonically increasing behavior for each session of the Oracle Cloud Agent software, resetting when the operating system is restarted.

2The Networking service provides more metrics (in the oci_vcn metric namespace) for each VNIC on the instance. For more information, see Networking Metrics.

3The Block Volume service provides more metrics (in the oci_blockstore metric namespace) for each volume attached to the instance. For more information, see Block Volume Metrics.

Using the Console

To view default metric charts for a single compute instance
  1. Open the navigation menu and click Compute. Under Compute, click Instances.
  2. Click the instance that you're interested in.
  3. Under Resources, click Metrics.
  4. In the Metric namespace list, select oci_computeagent.

    The Metrics page displays a default set of charts for the current instance.

    Not seeing any metric charts for the instance?

    If you don't see any metric charts, the instance might not be emitting metrics. See the following possible causes and resolutions.

    Possible cause How to check Resolution
    The Compute Instance Monitoring plugin is disabled on the instance or plugins are stopped. Review the instance properties. Enable the Compute Instance Monitoring plugin and start all plugins.
    The instance can't access the Monitoring service because its VCN doesn't use the internet. Review the instance's IP address. If it's not public, then a service gateway is needed. Set up a service gateway.
    The instance doesn't use a supported image. Review the supported images. Create an instance with a supported image.
    Older images and custom images: No Oracle Cloud Agent software exists on the instance. Connect to the instance and look for the software. Install the Oracle Cloud Agent software.
    Something else is wrong with the Oracle Cloud Agent software. (not applicable) Follow the troubleshooting steps for Oracle Cloud Agent.

    For more information about monitoring metrics and using alarms, see Monitoring. For information about notifications for alarms, see Notifications.

To view default metric charts for all compute instances in a compartment
  1. Open the navigation menu and click Observability & Management. Under Monitoring, click Service Metrics.
  2. Select a compartment.
  3. For Metric namespace, select oci_computeagent.

    The Service Metrics page dynamically updates the page to show charts for each metric that is emitted by the selected metric namespace.

For more information about monitoring metrics and using alarms, see Monitoring. For information about notifications for alarms, see Notifications.

To view default metric charts for the instances in an instance pool
  1. Open the navigation menu and click Compute. Under Compute, click Instance Pools.
  2. Click the instance pool that you're interested in.
  3. Under Resources, click Metrics.
  4. In the Metric namespace list, select oci_computeagent.

    The Metrics page displays a default set of charts for the current instance pool.

For more information about monitoring metrics and using alarms, see Monitoring. For information about notifications for alarms, see Notifications.