influxdb_grafana_poc

goal demonstrate & document a end2end monitoring flow using influxdb

what

• stat collection

  • use collectd on linux to sent stats to hekad
  • use http://ssc-serv.com/ which is collectd for windows

• stat aggregation

  • use hekad to aggregate stats & forward to influxdb
  • group agents in groups and aggregate certain stats e.g. IOPS over the groups & over all disks
    • create multiple levels of aggregation e.g. all disks in 1 machine for iops, all disks over all machines in 1 rack
    • so influxdb stores the aggregated values doesn't do the aggregation at query time

• stat visualization

  • create multiple dashboards in grafana to visualize stats

• docker

  • create docker images for the server side (start from despiegk/mc docker)
  • in the docker checkout this repo which has the required dashboards, hekascripts, ...
  • in the docker
    • influxdb
    • hekad (collectd sends to these)
    • collectd dumping to local hekad

• create howto docs

  • how to install collectd on each platform and configure to send to central hekad
  • how to get started with the docker image
  • explain with clear screenshots how to work with grafana & create new dashboards
  • explain how the hekad was configured and why

• what to visualize

  • iops per disk
  • disk space per disk (total,used)
  • network stats
  • ...

• snmp integration

  • install snmp agent (just to prove the point)
  • pull some info from the snmp agent and also dump to heka to get visualized later

installation

Installation