分布式监控(三)–采集器INPUT配置

2019独角兽企业重金招聘Python工程师标准>>>

1.操作系统基础监控指标配置标准

基础监控使用通用的全局配置文件telegraf.conf，以下只贴上采集器input部分代码

 

# Read metrics about cpu usage

[[inputs.cpu]]

  interval = “60s”

  percpu = true

  totalcpu = true

  fielddrop = [“time*”]

# Read metrics about disk usage by mount point

[[inputs.disk]]

  interval = “10m”

  # mount_points = [“/”]

  ignore_fs = [“tmpfs”, “devtmpfs”]

# Read metrics about disk IO by device

[[inputs.diskio]]

  interval = “60s”

# Get kernel statistics from /proc/stat

[[inputs.kernel]]

  interval = “10m”

# Read metrics about memory usage

[[inputs.mem]]

  interval = “10m”

# Get the number of processes and group them by status

[[inputs.processes]]

  interval = “10m”

  fielddrop = [“wait”,”idle”,”unknown”]

# Read metrics about swap memory usage

[[inputs.swap]]

  interval = “1h”

  fieldpass = [“used_percent”]

# Read metrics about system load & uptime

[[inputs.system]]

  interval = “2m”

# # Read metrics about network interface usage

 [[inputs.net]]

  interval = “60s”

  fieldpass = [“packets_*”,”bytes_*”,”drop_*”,”err_*”]

# # Read TCP metrics such as established, time wait and sockets counts.

 [[inputs.netstat]]

 interval = “5m”

 fielddrop = [“tcp_none”,”tcp_closing”,”tcp_close*”]

# # Read ping metrics

[[inputs.ping]]

 urls = [“www.qq.com”]

 count = 1

 ping_interval = 0.0

 timeout = 0.0

 interface = “eth0”

 

2.docker容器监控配置标准

对于基础监控之外的指标监控，需要使用新的配置文件，指定采集器的采集标准和配置，以及输出数据库（可选非必要），下同。

[[inputs.docker]]

   interval = “5m”

   endpoint = “unix:///var/run/docker.sock”

   container_names = []

   timeout = “10s”

   perdevice = true

   total = false

3.mysql监控配置标准

# Read metrics from one or many mysql servers

[[inputs.mysql]]

  interval = “5m”                  #mysql采集器频率

  servers = [“root:Root123@tcp(127.0.0.1:3306)/?tls=false”]          #采集器配置

  perf_events_statements_digest_text_limit  = 120

  perf_events_statements_limit              = 250

  perf_events_statements_time_limit         = 86400

  table_schema_databases                    = [“”]

  gather_table_schema                       = false

  

  gather_process_list                       = true

  

  gather_info_schema_auto_inc               = true

  

  gather_slave_status                       = true

  gather_binary_logs                        = false

  gather_table_io_waits                     = false

  gather_table_lock_waits                   = false

  gather_index_io_waits                     = false

  gather_event_waits                        = false

  gather_file_events_stats                  = false

  interval_slow                             = “30m”

 

4.haproxy配置标准

#vim /etc/telegraf/telegraf.d/mysql.conf

 

[inputs.haproxy]

        interval = “60s”                   #input采集频率

        servers = [“/var/run/haproxy.sock”]             #input参数配置

 

5.nginx配置标准

[[inputs.nginx]]

  interval = “60s”

  urls = [“http://localhost/server_status”]   

 

6.kafka配置标准

[[inputs.kafka_consumer]]

  topics = [“telegraf”]

  zookeeper_peers = [“localhost:2181”]

  consumer_group = “telegraf_metrics_consumers”

  metric_buffer = 100000

  offset = “oldest”

  data_format = “influx”

 

 

7.zookeeper配置标准

[[inputs.zookeeper]]

   servers = [“:2181”]

 

8.ovirt配置标准

9.redis配置标准

# # Read metrics from one or many redis servers

 [[inputs.redis]]

   interval = “5m”

   unix:///var/run/redis.sock

   servers = [“tcp://localhost:6379”]

 

10.http网站监控

[[inputs.http_response]]

 

 

      interval = “1m”

      address = “http://abc.com”

      response_timeout = “15s”

      method = “GET”

      follow_redirects = true

11.Ceph监控

[[inputs.ceph]]

  interval = ‘1m’

  ceph_binary = “/usr/bin/ceph”

  socket_dir = “/var/run/ceph”

  mon_prefix = “ceph-mon”

  osd_prefix = “ceph-osd”

  socket_suffix = “asok”

  ceph_user = “client.admin”

  ceph_config = “/etc/ceph/ceph.conf”

  gather_admin_socket_stats = true

  gather_cluster_stats = false

12.ping监控

单独使用场景：主要作为WAN链路监控，URL监控的监测手段

[[inputs.ping]]

     urls = [“www.qq.com”]

     count = 1

     ping_interval = 0.0

     timeout = 0.0

     interface = “eth0”

 

13.tcp端口监控

使用场景：主要作为本机端口应用监控

# # Generic TCP listener

[[inputs.tcp_listener]]

     service_address = “:8094”

     allowed_pending_messages = 10000

     max_tcp_connections = 250

     data_format = “influx”

14.端口流量监控

使用场景：主要作为入口或者外出服务器或者防火墙的端口流量监控

15.IPMI监控

使用场景：主要作为物理机器基础电力设施的指标监控

# #IPMI1# # 

[[inputs.ipmi_sensor]]

  servers = [“root:password@lanplus(10.8.81.15)”]

 

16.http-json格式监控

使用场景：在特殊应用无任何匹配模板时，可使用json格式化数据后进行指标收集；

# [[inputs.httpjson]]

   name = “webserver_stats”

   servers = [

     “http://localhost:9999/stats/”,

     “http://localhost:9998/stats/”,

                   ]

   method = “GET”

   tag_keys = [

  “my_tag_1”,

   “my_tag_2”

                 ]

 

 

 

17.自定义sh脚本监控

使用场景：在以上都没有可用的监控采集配置外，可以使用自定义脚本格式收集指标；

[[inputs.exec]]

    commands = [“/tmp/test.sh”,]

    timeout = “15s”

    data_format = “json”

    name_suffix = “_mycollector”

18.SQLserver监控

使用场景：用于监控SQLserver数据库性能

[[inputs.sqlserver]]

   servers = [

    “Server=192.168.1.30;Port=1433;User Id=telegraf;Password=T$l$gr@f69*;app name=dbname;log=1;”

    ]

转载于:https://my.oschina.net/kevin728971010/blog/1486119