chaosdb 0.2.12

A Chaostoolkit control to store events on a time series datastore or Grafana

chaosdb: Chaostoolkit storeevent Control

This control allows you to store events on a configurable data store. The data store can then be used to monitor events on a monitoring dashboard or to collect events on a time series datastore. The control currently implements the following drivers:

• an influx driver to store events on a InfluxDB time series database
• a grafana driver to store events on a Grafana server as annotations

The grafana driver supports the following authentication types:

• basic auth (username/password)
• bearer API token

To understand what a chaos-toolkit control is please refer to the official documentation on controls.

The Influx driver

InfluxDB is one of the most used time series data stores, it's implemented in golang and therefore is pretty fast and very easy to install and set up. Moreover, if you use Gatling for your performance input load simulations, it is pretty simple to store Gatling metrics on InfluxDB as well (as the Influx server can be configured to accept data in the graphite protocol), so you can use the same Influx data store for Gatling metrics and chaos toolkit events.

You can configure the Influx driver setting these parameters in the configuration section (the values provided represent the defaults):

    "configuration": {
      "influxdb": {
        "host": "localhost",
        "port": 8086,
        "http_endpoint": "/write",
        "database": "gatlingdb"
      }
    }

Then, at the proper level, configure the control driver:

            "controls": [
                {
                    "name": "tracing",
                    "provider": {
                        "type": "python",
                        "module": "chaosdb.influx"
                    }
                }
            ],

The Grafana driver

The Grafana driver can be quite convenient if you use grafana for your dashboards. The driver sends chaos-toolkit events directly to Grafana using the Annotation HTTP API. Moreover it is able to draw a region annotation around the whole experiment, making experiment visualization more visible.

The grafana configuration section is keyed "grafana". The grafana driver accepts a list of configuration sections, one for each grafana server you want to send annotations to. Each section accepts the following configuration attributes:

Attribute	Description
host	the grafana server fqdn
port	the grafana service port
protocol	http/https (defaults to http if not set)
api_token	grafana API token (not mandatory)
cert_file	file where the TSL certificate is stored (not mandatory)
username	the simple auth user name (not mandatory)
password	the simple auth password (not mandatory)
dashboardId	a specific dashboard ID (all in case it's not specified or 0)
only_actions	a boolean flag to send only actions or probes as well (defaults to false)
tags	a list of custom tags to tag annotations
annotation_api_endpoint	defaults to '/api/annotations'

The 'dashboardId' parameter points to the dashboard annotations are added to, otherwise annotations will be displayed on all dashboards.

The 'only_actions' parameter allows to trace only actions, not probes. The 'tags' parameter allows to add custom tags to each annotation.

Then, at the proper level, configure the control driver:

            "controls": [
                {
                    "name": "tracing",
                    "provider": {
                        "type": "python",
                        "module": "chaosdb.grafana"
                    }
                }
            ],