一、配置文件

进入到如下的配置文件

[root@localhost ~]# VIM /etc/telegraf/telegraf.conf

[global_tags]

[agent]

# 收集数据的间隔

interval = "10s"

round_interval = true

metric_batch_size = 1000

metric_buffer_limit = 10000

collection_jitter = "0s"

# 写入 inflxdb 的时间间隔

flush_interval = "10s"

flush_jitter = "0s"

precision = ""

debug = false

quiet = false

# 日志路径

logfile = "/data/log/telegraf.log"

hostname = ""

omit_hostname = false

[[outputs.influxdb]]

urls = ["http://192.168.227.56:8086"]

database = "telegraf"

username = "admin"

password = "123456"

[[inputs.cpu]]

# 是否采集每个 CPU 的数据

percpu = false

# 是否采集总 CPU 的数据

totalcpu = true

collect_cpu_time = false

report_active = false

二、启动 Telegraf

[root@localhost data]# systemctl start telegraf.service

[root@localhost data]# service telegraf status

开机启动

[root@localhost data]# systemctl enable telegraf.service

三、在 inflxdb 验证数据(inflxdb

数据库所在的机器)

[root@localhost ~]# influx -username 'admin' -password '123456'

-database telegraf

select * from CPU limit 3

name: CPU

time CPU host usage_guest usage_guest_nice usage_idle

usage_iowait usage_irq usage_nice usage_softirq usage_steal

usage_system usage_user

1598452090000000000 cpu-total localhost.localdomain 0 0

97.93294681121132 0.17645576002015695 0.15124779430301213 0

0.07562389715150607 0 0.8822788001008564 0.7814469372322055

1598452090000000000 cpu0 localhost.localdomain 0 0

98.58870967742497 0 0.20161290322583764 0 0 0 0.5040322580645762

0.7056451612904496

1598452090000000000 cpu1 localhost.localdomain 0 0

98.5901309164208 0 0.10070493454176956 0 0.10070493454178744 0

0.5035246727089193 0.70493454179253

四、在 grafana 展示监控 CPU 的数据

点击设置选择“data sources"

424 x 661

配置数据源

728 x 514 1154 x 814

728 x 230 1107 x 350

703 x 335

创建面板

694 x 271

点击 new dashboard 按钮

728 x 279 1411 x 541

点击 Graph 图标

728 x 493 732 x 496

点击“edit”按钮

728 x 375 780 x 402

选择数据源

728 x 411 886 x 500

添加需要关注的指标值：

728 x 350 1068 x 514

百分比占用显示：

728 x 586 752 x 605

五、选择保存默认项

[root@localhost telegraf]# VIM /etc/telegraf/telegraf.conf

对 CPU 只采集的监控项

fieldpass = ["usage_user","usage_idle", "usage_iowait",

"usage_system"]

重启 telegraf

[root@localhost telegraf]# systemctl restart telegraf

六、在 inflxdb 上面重新采集数据

[root@localhost ~]# influx -username 'admin' -password '123456'

-database telegraf

use telegraf;

drop measurement cpu;

728 x 338 1842 x 855

选择刷新频率和线上的时间段

作者：mike.liu

链接：http://testingpai.com/article/1601096205235

来源：测试社区http://testingpai.com