ping_exporter

Prometheus exporter for ICMP echo requests using https://github.com/digineo/go-ping

This is a simple server that scrapes go-ping stats and exports them via HTTP for Prometheus consumption. The go-ping library is build and maintained by Digineo GmbH. For more information check the source code.

Getting Started

Config file

Targets can be specified in a YAML based config file:

targets:
  - 8.8.8.8
  - 8.8.4.4
  - 2001:4860:4860::8888
  - 2001:4860:4860::8844
  - google.com:
      asn: 15169

dns:
  refresh: 2m15s
  nameserver: 1.1.1.1

ping:
  interval: 2s
  timeout: 3s
  history-size: 42
  size: 120

options:
  disableIPv6: false

Note: domains are resolved (regularly) to their corresponding A and AAAA records (IPv4 and IPv6). By default, ping_exporter uses the system resolver to translate domain names to IP addresses. You can override the resolver address by specifying the --dns.nameserver flag when starting the binary, e.g.

$ # use Cloudflare's public DNS server
$ ./ping_exporter --dns.nameserver=1.1.1.1:53 [other options]

Exported metrics

• ping_rtt_best_seconds: Best round trip time in seconds
• ping_rtt_worst_seconds: Worst round trip time in seconds
• ping_rtt_mean_seconds: Mean round trip time in seconds
• ping_rtt_std_deviation_seconds: Standard deviation in seconds
• ping_loss_ratio: Packet loss as a value from 0.0 to 1.0

Each metric has labels ip (the target's IP address), ip_version (4 or 6, corresponding to the IP version), and target (the target's name).

Additionally, a ping_up metric reports whether the exporter is running (and in which version).

Shell

To run the exporter:

$ ./ping_exporter [options] target1 target2 ...

or

$ ./ping_exporter --config.path my-config-file [options]

Help on flags:

$ ./ping_exporter --help

Getting the results for testing via cURL:

$ curl http://localhost:9427/metrics

Running as non-root user

On Linux systems CAP_NET_RAW is required to run ping_exporter as unpriviliged user.

# setcap cap_net_raw+ep /path/to/ping_exporter

When run through a rootless Docker implementation on Linux, the flag --cap-add=CAP_NET_RAW should be added to the docker run invocation.

If being invoked via systemd, you can alternately just add the following settings to the service's unit file in the [Service] section:

CapabilityBoundingSet=CAP_NET_RAW
AmbientCapabilities=CAP_NET_RAW

Docker

https://hub.docker.com/r/czerwonk/ping_exporter

To run the ping_exporter as a Docker container, run:

$ docker run -p 9427:9427 -v /path/to/config/directory:/config:ro --name ping_exporter czerwonk/ping_exporter

Kubernetes

To run the ping_exporter in Kubernetes, you can use the supplied helm chart

Prerequisites

• Helm v3.0.0+

Installing the chart

To install the chart with the release name ping-exporter:

$ helm repo add ping-exporter "https://raw.githubusercontent.com/czerwonk/ping_exporter/master/dist/charts/"
"ping-exporter" has been added to your repositories

$ helm repo update
Hang tight while we grab the latest from your chart repositories...
...Successfully got an update from the "ping-exporter" chart repository
Update Complete. ⎈Happy Helming!⎈

$ helm install ping-exporter ping-exporter/ping-exporter
NAME: ping-exporter
...

General parameters

Changes from previous versions

ping_loss_ratio vs ping_loss_percent

Previous versions of the exporter reported packet loss via a metric named ping_loss_percent. This was somewhat misleading / wrong, because it never actually reported a percent value (it was always a value between 0 and 1). To make this more clear, and to match with Prometheus best practices, this metric has been renamed to ping_loss_ratio instead.

If you had already been using an earlier version and want to continue to record this metric in Prometheus using the old name, this can be done using the metric_relabel_configs options in the Prometheus config, like so:

- job_name: "ping"
  static_configs:
    <...>
  metric_relabel_configs:
    - source_labels: [__name__]
      regex: "ping_loss_ratio"
      target_label: __name__
      replacement: "ping_loss_percent"

Time units

As per the recommendations for Prometheus best practices, the exporter reports time values in seconds by default. Previous versions defaulted to reporting milliseconds by default (with metric names ending in _ms instead of _seconds), so if you are upgrading from an older version, this may require some adjustment.

It is possible to change the ping exporter to report times in milliseconds instead (this is not recommended, but may be useful for compatibility with older versions, etc). To do this, the metrics.rttunit command-line switch can be used:

$ # keep using seconds (default)
$ ./ping_exporter --metrics.rttunit=s [other options]
$ # use milliseconds instead
$ ./ping_exporter --metrics.rttunit=ms [other options]
$ # report both millis and seconds
$ ./ping_exporter --metrics.rttunit=both [other options]

If you used the ping_exporter in the past, and want to migrate, start using --metrics.rttunit=both now. This gives you the opportunity to update all your alerts, dashboards, and other software depending on ms values to use proper scale (you "just" need to apply a factor of 1000 on everything). When you're ready, you just need to switch to --metrics.rttunit=s (or just remove the command-line option entirely).

Deprecated metrics

Previous versions of this exporter provided an older form of the RTT metrics as:

• ping_rtt_ms: Round trip times in millis

This metric had a label type with one of the following values:

• best denotes best round trip time
• worst denotes worst round trip time
• mean denotes mean round trip time
• std_dev denotes standard deviation

These metrics are no longer exported by default, but can be enabled for backwards compatibility using the --metrics.deprecated command-line flag:

$ # also export deprecated metrics
$ ./ping_exporter --metrics.deprecated=enable [other options]
$ # or omit deprecated metrics (default)
$ ./ping_exporter --metrics.deprecated=disable [other options]

Contribute

Simply fork and create a pull-request. We'll try to respond in a timely fashion.

License

MIT License, Copyright (c) 2018 Philip Berndroth pberndro Daniel Czerwonk dan_nrw