System metrics

更新时间:
复制 MD 格式

Use system metrics to monitor the resources of the virtual machine or ECS where an application runs. These resources include CPU, load, and memory. No log configuration is required. After you enable this feature, the monitoring agent automatically collects the data.

In the console, click Application Monitoring in the navigation pane on the left. Select My Applications or All Applications. Click an application to go to its details page. This topic describes the system metrics for application visualization. For more information about custom monitoring configurations, see Application Overview.

RMS supports the following system metrics:

Metric

Description

cpu_total_cores

The number of container cores.

cpu_util

CPU usage. Unit: %.

This metric consists of user+sys+nice+guest+hirq+sirq and reflects the current CPU utilization of the system. A higher percentage indicates a busier CPU.

cpu_sys

The percentage of time spent in kernel mode. Unit: %.

This typically refers to the time spent in kernel mode after an application invokes a syscall, such as for disk I/O writes, network communication, or other I/O operations. A high sys value may indicate kernel or hardware issues.

cpu_steal

The percentage of stolen CPU time. Unit: %.

cpu_wait

The percentage of time spent waiting for I/O. Unit: %.

cpu_user

The percentage of time spent in user mode. Unit: %.

load_load1

The load average over the last 1 minute.

load_load5

The load average over the last 5 minutes.

load_load15

The load average over the last 15 minutes.

mem_util

Memory usage. Unit: %.

mem_used

The total amount of memory used by users and the system.

This is the amount of memory calculated as total - buff - cache - free. Note that this does not include cache.

mem_total

The total amount of memory.

traffic_bytin

The inbound traffic rate of the network interface controller (NIC). Unit: bytes/s.

traffic_bytout

The outbound traffic rate of the NIC. Unit: bytes/s.

traffic_pktin

The inbound packets per second (pps) rate of the NIC.

traffic_pktout

The outbound pps rate of the NIC.

traffic_pktdrp

The total number of inbound and outbound dropped packets. Inside a container that uses a veth pair, this value is typically 0.

traffic_pkterr

The total number of inbound and outbound error packets. Inside a container that uses a veth pair, this value is typically 0.

tcp_active

Active opens per second. The number of times per second that connections are actively initiated.

tcp_lisove

Listen overflows. The number of connections dropped because the listen accept queue is full.

tcp_AtmpFail

AttemptFails. The number of failed connection attempts per second. For example, SYN packets are dropped because the half-open connection queue is full.

tcp_CurrEstab

Current Established connections. The current number of TCP connections.

tcp_pasive

Passive opens per second. The number of times per second that connections are passively generated.

tcp_retran

Retransmit ratio. The retransmission rate calculated from the number of retransmitted segments and the total number of segments.

udp_idgm

The number of UDP packets received by the host per second. Use this metric with `udp_odgm` to monitor the host's UDP message traffic.

udp_odgm

The number of UDP packets sent by the host per second.

udp_idmerr

The number of UDP error packets, including UDP checksum errors. This metric is typically used to monitor abnormal UDP communication.

udp_noport

The number of packets per second sent to an unknown UDP port. This occurs when no process on the host is listening on that port.

partition_util

The percentage of used disk space. Unit: %.

partition_bfree

The amount of free space. Unit: byte. The remaining available disk space on the mount target.

partition_bused

The amount of used space. Unit: byte. The amount of disk space that has been used.

partition_btotl

The total disk space. Unit: byte. The total available disk space on the mount target.

partition_ifree

The number of available inodes in the file system. The ext4 file system has a limit on the total number of inodes that can be used.

partition_itotl

The total number of available inodes.

partition_iutil

The percentage of used inodes out of the total available inodes. The overall usage rate. Unit: %.

io_wio

The number of write I/O operations per second.

io-rio

The number of read I/O operations per second.

io_rbytes

The number of bytes read per second.

io_wbytes

The number of bytes written per second.

io_rqsize

Request sector per I/O. The average size of each I/O request. Unit: KB.

io_qusize

Queue size. The current number of I/O requests in the I/O queue.

io_await

Average response time for I/O requests. The average time from when an I/O request is initiated to when it is completed. Unit: ms.

io_svctm

Service time. The time it takes for a request to be processed by the device, not including the time spent waiting in the queue.