https://github.com/m-lab/disco

DIScard COllector. Collects metrics from switches, writes them to disk and exports them.
https://github.com/m-lab/disco

Last synced: about 1 month ago
JSON representation

DIScard COllector. Collects metrics from switches, writes them to disk and exports them.

• Host: GitHub
• URL: https://github.com/m-lab/disco
• Owner: m-lab
• License: apache-2.0
• Created: 2020-06-15T19:14:31.000Z (over 4 years ago)
• Default Branch: main
• Last Pushed: 2024-06-10T03:05:30.000Z (7 months ago)
• Last Synced: 2024-10-29T15:59:58.568Z (about 2 months ago)
• Language: Go
• Size: 64.5 KB
• Stars: 1
• Watchers: 13
• Forks: 3
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
[![Coverage Status](https://coveralls.io/repos/github/m-lab/disco/badge.svg?branch=master)](https://coveralls.io/github/m-lab/disco?branch=master)

# DISCOv2

DISCOv2 is a reimplementation of the DISCO parts of

[collectd-mlab](https://github.com/m-lab/collectd-mlab), written in Go. It

differs a bit from DISCO, but fundamentally they are the same in that they

both probe a switch for various traffic metrics every 10s and then

periodically write the results to a file in JSON format. Pusher will upload

those files to GCS where they will be processed and parsed into BigQuery.

DISCOv2 supports the following flags:

* `--prometheusx.listen-address`: the IP and TCP port to listen to for

   Prometheus metricis requests.

* `--metrics-file`: the path to a YAML-formatted file defining which metrics to

   scrape. See file metrics.yaml in this repo for an example.

* `--write-interval`: the interval at which collected metrics are converted to

   JSON and written to disk.

* `--target`: the name or IP of the switch to collect metrics from.

DISCOv2 requires that an environment variable named `DISCO_COMMUNITY` is set

and contains the SNMP community sting to use when polling the switch.

Unlike DISCO, in addition to collecting switch metrics every 10s and writing

out data files, DISCOv2 includes a Prometheus exporter which will expose the

metrics it has collected. This makes DISCOv2 something like the

[snmp_exporter](https://github.com/prometheus/snmp_exporter), but far less

general purpose.

# Configuration

The file metrics.yaml.sample is, as the name implies, nothing more than a sample

of how the configuration file should be formatted. The [actual configuration

file for

DISCOv2](https://github.com/m-lab/k8s-support/blob/master/config/disco/metrics.yaml)

(as it runs in the M-Lab kubernetes platform cluster) can be found in the

k8s-support repository.