OCI GPU

Supported OS Linux

This page is not yet available in Spanish. We are working on its translation.
If you have any questions or feedback about our current translation project, feel free to reach out to us!

Overview

Monitoring Oracle Cloud Infrastructure (OCI) GPU instances is essential for ensuring optimal performance and reliability of your high-performance computing workloads. This integration provides a comprehensive set of GPU metrics through the gpu_infrastructure_health namespace, enabling you to track various aspects of GPU health and utilization.

This integration lets you monitor and alert on the health, capacity, throughput, status, and performance of your GPU Instances.

It collects metrics and tags from the gpu_infrastructure_health namespace.

Setup

After setting up the Oracle Cloud Infrastructure integration, ensure that any namespaces mentioned above are included in your Connector Hub.

Data Collected

Metrics

oci.gpu_infrastructure_health.gpu_ecc_double_bit_errors
(count)		The number of GPU double-bit ECC errors reported.
Shown as error
oci.gpu_infrastructure_health.gpu_ecc_single_bit_errors
(count)		The number of GPU single-bit ECC errors reported.
Shown as error
oci.gpu_infrastructure_health.gpu_memory_utilization
(gauge)		The percentage of the GPU memory resource in use.
Shown as percent
oci.gpu_infrastructure_health.gpu_power_draw
(gauge)		The amount of GPU power used.
oci.gpu_infrastructure_health.gpu_temperature
(gauge)		The GPU temperature reported.
oci.gpu_infrastructure_health.gpu_utilization
(gauge)		Activity level from GPU. Expressed as a percentage of total time. For instance pools, the value is averaged across all instances in the pool.
Shown as percent

Service Checks

OCI GPU does not include any service checks.

Events

OCI GPU does not include any events.

Troubleshooting

Need help? Contact Datadog support.