Monitoring an OpenStack environment becomes a more and more complex task. This talk will keep an eye especially on performance, mostly ignoring availability checks.
There are a few projects specialized on collecting performance metrics from systems.
OpenStack deployments can also grow into bigger sizes, where nodes are unavailable or dynamicly added at run time. During the talk, we'll evaluate several projects:
- maybe: metricd (if time permits)
At the end, we'll focus on a solution using collectd, gnocchi with ceph backed and grafana for visualization.
Attendees should get an overview on current available Open Source solutions for monitoring performance. A selection will be investigated and at the end, one selection of projects will be described in depth.