####configuration overview

The goal of this playbook is to make use of InfluxDB as a system metric store. Features:

• Can automatically cluster n-number of influxdb nodes
• Uses Telegraf to gather metrics using UDP
• Sets up Kapacitor for processing, monitoring, and alerting
• Sets up a grafana dashboard
• Sets up downsampling for all metrics

####notes

• All of the packages in this repo are set to state=latest so that I don't have to continuously maintain the version pin. If you intend to use this in production you should pin all of the versions and do testing between upgrades. Additionally the influxdb modules make a lot of assumptions about the way influx does things and are vulnerable to being broken between versions.

####requirements

• Virtualbox
• Vagrant
• Ansible via pip install ansible

####running in vagrant

vagrant up
bin/provision_development

####destroying

bin/destroy

####influxdb admin site

http://10.0.168.101:8083

####grafana dashboard

http://10.0.168.101:3006

####custom ansible modules

• influxdb_user - create users for Influxdb
• influxdb_retention_policy - create retention policies for Influxdb
• influxdb_downsample - given measurements [{'measurement':'system', 'fields':['load1','load5','load15']] and retention policy 30s:30d,1m:52w continuous queries will be generated that downsample the metrics accordingly. The naming scheme will generate the continuous query system_30s with assigned retention policy rp_30d.

####custom filter plugins

• to_url_with_port - assist in building the list of influxdb servers for telegraf.
  • Example: hostvars|to_url_with_port(groups,'influxdb','udp://','private_ip','8089')|join(', ')