Good morning everybody!
Guys, which monitoring stack do you recommend to monitor Host and VM resources? Or what monitoring strategies do they recommend?
Ceilometer, specially with the sub-system of dynamic pollsters. Then, you can leverage Gnocchi as a storage backend for the monitored metrics.
On Thu, Jun 8, 2023 at 9:25 AM Murilo Morais murilo@evocorp.com.br wrote:
Good morning everybody!
Guys, which monitoring stack do you recommend to monitor Host and VM resources? Or what monitoring strategies do they recommend?
On Thu, Jun 08, 2023 at 09:19:11AM -0300, Murilo Morais wrote:
Good morning everybody!
Good morning,
Guys, which monitoring stack do you recommend to monitor Host and VM resources? Or what monitoring strategies do they recommend?
This highly depends on your understanding of the term "monitoring".
To track resources used by OpenStack users/projects, I'd recommend ceilometer. In addition to ceilometer, you'll probably need something to track resource usage in infrastructure nodes, e.g memory usage, disk usage, I/O etc. My recommendation would be to use something like collectd or node_exporter.
Then you'd need a data storage for storing the metrics. Things like gnocchi or prometheus come into mind.
The choice of the tools will probably be influenced by the available knowledge, the way and where OpenStack is installed.
What is not covered above is: tracking if services are up and running, any kind of centralized logging, etc.
Matthias
Thanks a lot for the tips, I'll try it!
Em sex., 9 de jun. de 2023 às 03:00, Matthias Runge < mrunge@matthias-runge.de> escreveu:
On Thu, Jun 08, 2023 at 09:19:11AM -0300, Murilo Morais wrote:
Good morning everybody!
Good morning,
Guys, which monitoring stack do you recommend to monitor Host and VM resources? Or what monitoring strategies do they recommend?
This highly depends on your understanding of the term "monitoring".
To track resources used by OpenStack users/projects, I'd recommend ceilometer. In addition to ceilometer, you'll probably need something to track resource usage in infrastructure nodes, e.g memory usage, disk usage, I/O etc. My recommendation would be to use something like collectd or node_exporter.
Then you'd need a data storage for storing the metrics. Things like gnocchi or prometheus come into mind.
The choice of the tools will probably be influenced by the available knowledge, the way and where OpenStack is installed.
What is not covered above is: tracking if services are up and running, any kind of centralized logging, etc.
Matthias
-- Matthias Runge mrunge@matthias-runge.de
participants (3)
-
Matthias Runge
-
Murilo Morais
-
Rafael Weingärtner