• Cloud Eye

ces
  1. Help Center
  2. Cloud Eye
  3. User Guide
  4. Services Supported by Cloud Eye
  5. Computing
  6. ECS Metrics

ECS Metrics

ECS metrics vary depending on ECS OSs and types. For details, see Table 1. √ indicates that the metric is supported, and x indicates that the metric is not supported.

Table 1 ECS metrics

Metric

Windows ECS

Linux ECS

N/A

XEN

KVM

XEN

KVM

CPU Usage

Memory Usage

√ (The OTC Tools must be installed on the image. Otherwise, this metric is unavailable.)

x (Not supported)

Disk Usage

√ (The OTC Tools must be installed on the image. Otherwise, this metric is unavailable.)

x (Not supported)

Disk Read Bandwidth

Disk Write Bandwidth

Disk Read IOPS

Disk Write IOPS

Inband Incoming Rate

x (Not supported)

Inband Outgoing Rate

x (Not supported)

Outband Incoming Rate

Outband Outgoing Rate

System Status Check Failed

×

×

InfiniBand NIC status

×

√ (only for H2 ECSs)

×

√ (only for H2 ECSs)

NOTE:

For Memory Usage and Disk Usage, the OTC Tools must be installed for the images used by ECSs. Otherwise, the metrics cannot be obtained. For details about how to install the OTC Tools, visit https://github.com/UVP-Tools/UVP-Tools/.

Table 2 describes these ECS metrics.

Table 2 ECS metrics

Metric

Description

Formula

Remarks

CPU Usage

Indicates the vCPU usage (%) of an ECS.

vCPU usage of an ECS/Number of vCPUs in the ECS

N/A

Memory Usage

Indicates the memory usage (%) of an ECS.

Used memory of an ECS/Total memory of the ECS

This metric is unavailable if the image has no OTC Tools installed.

Disk Usage

Indicates the disk usage (%) of an ECS.

Used capacity of an ECS disk/Total capacity of the ECS disk

This metric is unavailable if the image has no OTC Tools installed.

Disk Read Bandwidth

Indicates the number of bytes read from an ECS per second.

Total number of bytes read from an ECS disk/Monitoring period

byte_out = (rd_bytes - last_rd_bytes) /Time difference

Disk Write Bandwidth

Indicates the number of bytes written to an ECS per second.

Total number of bytes written to an ECS disk/Monitoring period

N/A

Disk Read IOPS

Indicates the number of read requests sent to an ECS per second.

Total number of read requests sent to an ECS disk/Monitoring period

req_out = (rd_req - last_rd_req)/Time difference

Disk Write IOPS

Indicates the number of write requests sent to an ECS per second.

Total number of write requests sent to an ECS disk/Monitoring period

req_in = (wr_req - last_wr_req)/Time difference

Inband Incoming Rate

Indicates the number of incoming bytes on an ECS per second.

Total number of inband incoming bytes on an ECS/Monitoring period

N/A

Inband Outgoing Rate

Indicates the number of outgoing bytes on an ECS per second.

Total number of inband outgoing bytes on an ECS/Monitoring period

N/A

Outband Incoming Rate

Indicates the number of incoming bytes on an ECS per second on the hypervisor.

Total number of outband incoming bytes on an ECS/Monitoring period

This metric is unavailable if SR-IOV is enabled.

Outband Outgoing Rate

Indicates the number of outgoing bytes on an ECS per second on the hypervisor.

Total number of outband outgoing bytes on an ECS/Monitoring period

This metric is unavailable if SR-IOV is enabled.

System Status Check Failed

Indicates the status of the cloud platform for running ECSs. The check result is 0 or 1.

  • 0: The system is running properly. All check items are normal.
  • 1: The system is not running properly. At least one check item failed.

The system periodically checks the status and returns check results using value 0 or 1.

  • If the detected fault does not affect ECS functions, certain management operations performed on the ECS, such as starting, stopping, or specifications modifications, may be affected.
  • If the detected fault affects ECS functions, such as a host power supply failure, the system will recover the ECS as soon as possible.

When the power source of the physical host fails or the hardware/software becomes faulty, the check result is 1.

InfiniBand NIC status

Indicates the status of an InfiniBand NIC on an H2 ECS.

NOTE:

Only Mellanox EDR 100 GB single-port InfiniBand NICs are supported.

The system periodically checks the status and returns check results using value 0 or 1.

  • 0: The system is running properly. That is, the InfiniBand NIC is functional.
  • 1: The system is not running properly. That is, the InfiniBand NIC malfunctions.

When the physical NIC corresponding to a virtual NIC becomes faulty, for example, the network cable is not securely connected to the NIC, the switch or adapter is incompatible with the InifiniBand NIC, or the NIC is disabled, the returned value is 1.