telegraf数据采集端部署与使用

Linux大区 1年前 (2022) 导航君

27 0 0

环境

-夜莺5.1

telegraf-1.21.0

telegraf安装与基础配置

centos7下安装telegraf采集端

mkdir -p /opt/telegraf
wget  https://dl.influxdata.com/telegraf/releases/telegraf-1.21.0_linux_amd64.tar.gz
tar xf telegraf-1.21.0_linux_amd64.tar.gz 
cp  telegraf-1.21.0/usr/bin/telegraf /opt/telegraf/
cat  > /opt/telegraf/telegraf.conf <<EOF
[global_tags]
  #全局标签
  platform = "88lulu"
[agent]
  ## 默认数据采集间隔10s一次，这里1分钟一次，全局的，可以在每个模块上使用interval
  interval = "60s"
  round_interval = true
  metric_batch_size = 1000
  metric_buffer_limit = 10000
  collection_jitter = "0s"
  flush_interval = "10s"
  flush_jitter = "0s"
  precision = ""
  ## 默认主机名
  hostname = ""
  omit_hostname = false
[[outputs.opentsdb]]
  ##配置采集数据上传到夜莺上
  host = "http://192.168.31.9"
  port = 19000
  http_batch_size = 50
  http_path = "/opentsdb/put"
  debug = false
  separator = "_"

[[inputs.cpu]]
  percpu = true
  totalcpu = true
  collect_cpu_time = false
  report_active = true

[[inputs.disk]]
  ##采集对应的挂载点
  mount_points = ["/","/data1","/data2"]
  ignore_fs = ["tmpfs", "devtmpfs", "devfs", "iso9660", "overlay", "aufs", "squashfs"]
[[inputs.diskio]]
  ##采集对应的磁盘类型
  devices = ["sda", "sdb", "vd*"]
[[inputs.kernel]]

[[inputs.mem]]
  ## 收集需要匹配的项，根据自己需求收集
  fieldpass = ["*available", "*total","*percent","*used"]

[[inputs.swap]]
  fieldpass = ["*used","*total", "*percent"]

[[inputs.processes]]

[[inputs.system]]
  fielddrop = ["uptime_format"]

[[inputs.net]]
  ## 收集指定网卡eth0
  interfaces = ["eth0"]
  ignore_protocol_stats =  true
[[inputs.netstat]]
EOF

cat  > /etc/systemd/system/telegraf.service <<EOF
[Unit]
Description="telegraf"
After=network.target
[Service]
Type=simple
ExecStart=/opt/telegraf/telegraf --config telegraf.conf
WorkingDirectory=/opt/telegraf
Restart=on-failure
SuccessExitStatus=0
LimitNOFILE=65536
StandardOutput=syslog
StandardError=syslog
SyslogIdentifier=telegraf
[Install]
WantedBy=multi-user.target
EOF
systemctl daemon-reload
systemctl enable telegraf
systemctl restart telegraf
systemctl status telegraf

centos6下安装telegraf采集端

```shell
wget https://dl.influxdata.com/telegraf/releases/telegraf-1.21.1-1.x86_64.rpm
yum install telegraf-1.21.1-1.x86_64.rpm -y
cat  > /etc/telegraf/telegraf.conf<<EOF
[global_tags]
  #全局标签
  platform = "88lulu"
[agent]
  ## 默认数据采集间隔10s一次，这里1分钟一次，全局的，可以在每个模块上使用interval
  interval = "60s"
  round_interval = true
  metric_batch_size = 1000
  metric_buffer_limit = 10000
  collection_jitter = "0s"
  flush_interval = "10s"
  flush_jitter = "0s"
  precision = ""
  ## 默认主机名
  hostname = ""
  omit_hostname = false
[[outputs.opentsdb]]
  ##配置采集数据上传到夜莺上
  host = "http://192.168.31.9"
  port = 19000
  http_batch_size = 50
  http_path = "/opentsdb/put"
  debug = false
  separator = "_"

[[inputs.cpu]]
  percpu = true
  totalcpu = true
  collect_cpu_time = false
  report_active = true

[[inputs.disk]]
  ##采集对应的挂载点
  mount_points = ["/","/data1","/data2"]
  ignore_fs = ["tmpfs", "devtmpfs", "devfs", "iso9660", "overlay", "aufs", "squashfs"]
[[inputs.diskio]]
  ##采集对应的磁盘类型
  devices = ["sda", "sdb", "vd*"]
[[inputs.kernel]]

[[inputs.mem]]
  ## 收集需要匹配的项，根据自己需求收集
  fieldpass = ["*available", "*total","*percent","*used"]

[[inputs.swap]]
  fieldpass = ["*used","*total", "*percent"]

[[inputs.processes]]

[[inputs.system]]
  fielddrop = ["uptime_format"]

[[inputs.net]]
  ## 收集指定网卡eth0
  interfaces = ["eth0"]
  ignore_protocol_stats =  true
[[inputs.netstat]]
EOF
chkconfig --add telegraf
/etc/init.d/telegraf  restart

telegraf基础数据采集项说明

cpu：[[inputs.cpu]]
cpu_usage_active cpu使用率%
内存：[[inputs.mem]]
mem_used_percent 内存使用率%
mem_available 内存剩余大小
mem_total 内存总大小
swap分区：[[inputs.swap]]
swap_used_percent swap分区使用率
磁盘：[[inputs.diskio]]
disk_used_percent 磁盘使用率
disk_free 磁盘剩余大小
disk_total 磁盘总大小
(disk_inodes_used/disk_inodes_total)*100 磁盘inodes使用率
rate(diskio_io_time[1m])/10 磁盘IO使用率
系统：[[inputs.system]]
system_load1 1分钟负载
system_load5 5分钟负载
system_load15 15分钟负载
进程：[[inputs.processes]]
processes_total 进程总数
processes_zombies 假死进程数
网络: [[inputs.net]]
rate(net_bytes_recv[5m]) 接收/进站流量，5分钟内,每秒的速率单位bytes/秒
rate(net_bytes_sent[5m]) 发送/出站流量，5分钟内,每秒的速率单位bytes/秒
TCP连接数：[[inputs.netstat]]
netstat_tcp_established 已建立的TCP连接数
netstat_tcp_close_wait 等待关闭的TCP连接数

telegraf的mysql数据采集说明

mysql模块

[[inputs.mysql]]
##mysql账号密码连--接
servers = ["root:123456@tcp(127.0.0.1:3306)/?tls=false"]
## telegraf版本>1.6 用2，否则1
metric_version = 2
# 默认收集所有库，否则需要指定库
table_schema_databases = []
# 从 INFORMATION_SCHEMA的PROCESSLIST表上 收集线程状态计数
gather_process_list = true
# 从信息模式中收集 auto_increment 列和最大值
gather_info_schema_auto_inc = true
# 从 SHOW SLAVE STATUS 命令输出中收集指标
gather_slave_status = true
# 从 SHOW GLOBAL VARIABLES 命令输出收集指标
gather_global_variables = true
perf_events_statements_digest_text_limit = 120
perf_events_statements_limit = 250
perf_events_statements_time_limit = 86400
interval_slow = "30m"

mysql指标说明(仅供参考)：

rate(mysql_com_show_processlist[5m]) * 60   5分钟内，每分钟mysql进程总数
rate(mysql_bytes_sent[5m])    5分钟内，每秒出流量单位bytes/每秒
rate(mysql_bytes_received[5m])  5分钟内，每秒入流量单位bytes/每秒
rate(mysql_aborted_connects[5m]) * 60   5分钟内，每分钟mysql连接错误数
rate(mysql_connections[5m]) * 60        5分钟内，每分钟mysql连接总数（包括错误连接数）
mysql_threads_connected                 当前打开的mysql连接数
rate(mysql_com_select[5m])    5分钟内，每秒查询的次数(读IOPS)
rate(mysql_com_insert[5m])+rate(mysql_com_update[5m])+rate(mysql_com_delete[5m])  5分钟内，每秒写入的次数(写IOPS)