Open leon-sagess opened 7 years ago
Now, looking at the data in Cassandra, I realize that the actual cpu value and io_time values are not stored. Any reason why?
Could you elaborate a bit? Usually they should be stored. Problem is that we reduce it to the mean.
Usually you have to get a dropdown where you could both fetch _sum
, _mean
_max
and _mean
..
Thanks for commenting. Indeed I get those suggestions in the dropdown. But no matter what i select, the graph's query is set to _sum. See attached screenshot. My query is idle_mean. But the graph is still idle_sum.
Oh, this is a bug. Thanks for the screenshot. I'll look into it (unless someone else wants to)
I would be very happy to test.
Hi Alex, when do you think you would have a chance to look at this? Thx.
I hope to do it this week, trying to find time.
Thanks. I am looking forward to it. I have some pressure to demo cyanite.
I've tried several times, but so far I could not reproduce the issue. All the metrics are displayed with their correspondent counterpart. I'm afraid until we have a proper binary release it will be hard or impossible to make sure everyone's using a correct version...
I have cyanite setup and I send metrics to it using collectd. From a setup perspective everything seems to work. My collectd config is attached.
The data in cassandra looks like this:
collectd.zip
Now, my issue is that any query such as cpu.idle ends up being ignored by grafana. Instead the query it ends up graphing is on cpu.idle_sum.
Similarly, if I create a query on disk.io_time, the graph automatically appends _sum to it and it creates a graph for disk.io_time_sum.
Now, looking at the data in Cassandra, I realize that the actual cpu value and io_time values are not stored. Any reason why?
On the other had, the mean and min values are stored in cassandra. But still if I create a query disk.io_time_mean, then graph still displays a graph for disk.io_time_sum.
What am I missing? How do I create a graph that shows me the cpu progressing over time?