Updated on 2024-08-16 GMT+08:00

Host View

The host view displays information about host resource usage and health status, as well as common metrics of system devices such as disks and file systems, so you can better monitor a node.

Metric Description

You can view the host metrics in the following table.

Figure 1 Host metrics
Table 1 Host metrics

Metric

Unit

Description

CPU Usage

%

Usage of each CPU core.

Load Average

N/A

CPU contention.

  • If the value is less than 1, some CPU resources are used for processing requests.
  • If the value is 1, all CPU resources are used for processing requests.
  • If the value is greater than 1, some threads are waiting for processing.

Used Memory

byte

Memory usage in each mode.

Memory Usage

%

Host memory usage.

Disk Written

byte/s

Write rate of each disk.

Disk Read

byte/s

Read rate of each disk.

Used Disk Space

byte

Used and available disk space.

Disk Space Usage

%

Disk usage of each device.

Disk I/O Latency

second

Disk I/O read/write latency, in seconds.

TCP Connection

N/A

The number of TCP sockets.

  • alloc: the number of TCP sockets allocated (established, and sk_buff obtained).
  • inuse: the number of TCP sockets in use (listening)
  • orphan: the number of TCP sockets not associated with applications.
  • tw: the number of TCP sockets waiting to be closed.

UDP Usage

N/A

UDP usage.

  • liteInuse: the number of UDP-Lite sockets in use.
  • inuse: the number of UDP sockets in use.
  • useMemory: the usage of the UPD buffer.

File Descriptor

exabyte

Maximum number of file descriptors.

Used File Descriptor

N/A

The number of allocated file descriptors.

Socket Usage

N/A

Socket usage.

  • socketsUsed: the total number of used protocol sockets.
  • fragInuse: the number of fragment sockets in use.
  • fragMemroy: the usage of the fragment buffer.
  • rawInuse: the number of raw sockets in use.

Abnormal File System

N/A

File system status.

  • readonly: The file system is read-only.
  • deviceError: A file system error occurred.

Disk I/O Rate

N/A

The number of read/write operations on a disk per second.

Disk Read/Write Latency

second

Disk read/write latency.

I/O Queues

N/A

Average I/O queue length of the disk device, which is the weighted number of seconds spent doing I/O operations. A larger value indicates better disk performance of the node.

Process State

N/A

The number of processes in each state.

Connection Tracking Table Entries

N/A

  • Allocated: the number of allocated entries in the connection tracking table.
  • Total: the maximum number of entries in the connection tracking table.

Metric List

The following is the metric list of the host view.
Table 2 Metric description

Metric

Type

Description

node_cpu_seconds_total

Counter

Seconds the CPUs spent in each mode.

node_load1

Gauge

Average CPU load within 1 minute, which reflects the competition of CPU resources.

  • If the value is less than 1, some CPU resources are processing requests.
  • If the value is 1, all CPU resources are processing requests.
  • If the value is greater than 1, some threads are waiting for processing.

node_load15

Gauge

Average CPU load within 15 minutes.

node_memory_MemTotal_bytes

Gauge

Total memory of a node.

node_memory_MemAvailable_bytes

Gauge

Available memory of a node.

node_disk_written_bytes_total

Gauge

Total number of bytes written successfully.

node_disk_read_bytes_total

Gauge

Total number of bytes read successfully.

node_filesystem_size_bytes

Gauge

File system size in bytes.

node_filesystem_avail_bytes

Gauge

Available file system size in bytes.

node_disk_io_time_seconds_total

Counter

Total seconds spent doing I/O operations.

node_sockstat_TCP_alloc

Gauge

The number of allocated TCP sockets.

node_sockstat_UDPLITE_inuse

Gauge

The number of used UDPLITE sockets.

node_filefd_maximum

Gauge

File descriptor statistics: maximum.

node_filefd_allocated

Gauge

File descriptor statistics: allocated.

node_sockstat_sockets_used

Gauge

The number of used IPv4 sockets.

node_filesystem_readonly

Gauge

Read-only status of the file system.

node_disk_reads_completed_total

Counter

The number of disk reads completed.

node_disk_read_time_seconds_total

Counter

Total number of seconds spent by all reads.

node_disk_io_time_weighted_seconds_total

Counter

The weighted number of seconds spent doing I/O operations. A larger value indicates better disk performance of the node.

node_procs_blocked

Gauge

The number of processes blocked waiting for I/O to complete.

node_nf_conntrack_entries

Gauge

The number of currently allocated flow entries for connection tracking.