telegraf/plugins/inputs/ping
Daniel Nelson 63b047c91a
Fix ping skips remaining hosts after dns lookup error (#6743)
2019-12-03 11:27:33 -08:00
..
README.md Update ping input readme (#6412) 2019-09-20 16:49:14 -07:00
ping.go Fix ping skips remaining hosts after dns lookup error (#6743) 2019-12-03 11:27:33 -08:00
ping_notwindows.go Add native Go ping method to ping input plugin (#6050) 2019-07-11 15:07:58 -07:00
ping_test.go Fix incorrect results in ping plugin (#6581) 2019-11-05 10:34:18 -08:00
ping_windows.go Add native Go ping method to ping input plugin (#6050) 2019-07-11 15:07:58 -07:00
ping_windows_test.go

README.md

Ping Input Plugin

Sends a ping message by executing the system ping command and reports the results.

This plugin has two main methods of operation: exec and native. The recommended method is native, which has greater system compatibility and performance. However, for backwards compatibility the exec method is the default.

When using method = "exec", the systems ping utility is executed to send the ping packets.

Most ping command implementations are supported, one notable exception being that there is currently no support for GNU Inetutils ping. You may instead use the iputils-ping implementation:

apt-get install iputils-ping

When using method = "native" a ping is sent and the results are reported in native Go by the Telegraf process, eliminating the need to execute the system ping command.

Configuration:

[[inputs.ping]]
  ## Hosts to send ping packets to.
  urls = ["example.org"]

  ## Method used for sending pings, can be either "exec" or "native".  When set
  ## to "exec" the systems ping command will be executed.  When set to "native"
  ## the plugin will send pings directly.
  ##
  ## While the default is "exec" for backwards compatibility, new deployments
  ## are encouraged to use the "native" method for improved compatibility and
  ## performance.
  # method = "exec"

  ## Number of ping packets to send per interval.  Corresponds to the "-c"
  ## option of the ping command.
  # count = 1

  ## Time to wait between sending ping packets in seconds.  Operates like the
  ## "-i" option of the ping command.
  # ping_interval = 1.0

  ## If set, the time to wait for a ping response in seconds.  Operates like
  ## the "-W" option of the ping command.
  # timeout = 1.0

  ## If set, the total ping deadline, in seconds.  Operates like the -w option
  ## of the ping command.
  # deadline = 10

  ## Interface or source address to send ping from.  Operates like the -I or -S
  ## option of the ping command.
  # interface = ""

  ## Specify the ping executable binary.
  # binary = "ping"

  ## Arguments for ping command. When arguments is not empty, the command from
  ## the binary option will be used and other options (ping_interval, timeout,
  ## etc) will be ignored.
  # arguments = ["-c", "3"]

  ## Use only IPv6 addresses when resolving a hostname.
  # ipv6 = false

File Limit

Since this plugin runs the ping command, it may need to open multiple files per host. The number of files used is lessened with the native option but still many files are used. With a large host list you may receive a too many open files error.

To increase this limit on platforms using systemd the recommended method is to use the "drop-in directory", usually located at /etc/systemd/system/telegraf.service.d.

You can create or edit a drop-in file in the correct location using:

$ systemctl edit telegraf

Increase the number of open files:

[Service]
LimitNOFILE=8192

Restart Telegraf:

$ systemctl edit telegraf

Linux Permissions

When using method = "native", Telegraf will attempt to use privileged raw ICMP sockets. On most systems, doing so requires CAP_NET_RAW capabilities.

With systemd:

$ systemctl edit telegraf
[Service]
CapabilityBoundingSet=CAP_NET_RAW
AmbientCapabilities=CAP_NET_RAW
$ systemctl restart telegraf

Without systemd:

$ setcap cap_net_raw=eip /usr/bin/telegraf

Reference man 7 capabilities for more information about setting capabilities.

When Telegraf cannot listen on a privileged ICMP socket it will attempt to use ICMP echo sockets. If you wish to use this method you must ensure Telegraf's group, usually telegraf, is allowed to use ICMP echo sockets:

$ sysctl -w net.ipv4.ping_group_range="GROUP_ID_LOW   GROUP_ID_HIGH"

Reference man 7 icmp for more information about ICMP echo sockets and the ping_group_range setting.

Metrics

  • ping
    • tags:
      • url
    • fields:
      • packets_transmitted (integer)
      • packets_received (integer)
      • percent_packets_loss (float)
      • ttl (integer, Not available on Windows)
      • average_response_ms (integer)
      • minimum_response_ms (integer)
      • maximum_response_ms (integer)
      • standard_deviation_ms (integer, Available on Windows only with native ping)
      • errors (float, Windows only)
      • reply_received (integer, Windows with method = "exec" only)
      • percent_reply_loss (float, Windows with method = "exec" only)
      • result_code (int, success = 0, no such host = 1, ping error = 2)
reply_received vs packets_received

On Windows systems with method = "exec", the "Destination net unreachable" reply will increment packets_received but not reply_received*.

ttl

There is currently no support for TTL on windows with "native"; track progress at https://github.com/golang/go/issues/7175 and https://github.com/golang/go/issues/7174

Example Output

ping,url=example.org average_response_ms=23.066,ttl=63,maximum_response_ms=24.64,minimum_response_ms=22.451,packets_received=5i,packets_transmitted=5i,percent_packet_loss=0,result_code=0i,standard_deviation_ms=0.809 1535747258000000000