Hi Matthew!
We've been testing tach and it looks like it could be of great use for monitoring compute and controller nodes performance. I was wondering if gathering metrics from multiple nodes at the same time is supported, i.e. load balanced nova controller nodes for example.
Looking at the graphs in graphite I'd say that If you send metrics from multiple nodes at the same time they'll end up in the same folder. is that right? If so, how about prefixing the metric name with the hostname so you can get something like "timers.HOSTNAME.nova.api.foo.bar.metric"?
Thanks.
Hi Matthew!
We've been testing tach and it looks like it could be of great use for monitoring compute and controller nodes performance. I was wondering if gathering metrics from multiple nodes at the same time is supported, i.e. load balanced nova controller nodes for example.
Looking at the graphs in graphite I'd say that If you send metrics from multiple nodes at the same time they'll end up in the same folder. is that right? If so, how about prefixing the metric name with the hostname so you can get something like "timers.HOSTNAME.nova.api.foo.bar.metric"?
Thanks.