Hi,
I'm starting to mess with the various pcp-webjs options and have a
few questions on using the grafana component. cloned the github tree
and stuck it in /usr/share/pcp/webapps, started up the pmwebd service
and all is working fine. All very impressive actually! It picks up my
local archives, I can add metrics, add charts, etc.
A couple of questions:
1.
Not sure what they are called, but for lack of a better name I
can't get the "metric operations" to work: when you click the "plus
sign" and get the "Combine, Transform, Calculate....." options. I tried
a bunch of different combinations, but as a simple contrived example:
1. Edit the "1 minute load average"
2. Change "1 minute" to "*"
This works, shows 3 plots on the 1 graph
3. Click "+" and add Combine->maxSeries()
Results in "No Datapoints"
Doing the same sort of thing with the examples here seems to work:
http://play.grafana.org/
My real use case for this would be to load up "kernel.all.cpu.user" for
~100 hosts over a day and plot min,max,mean during the time period.
2.
My test case is about 1 month worth of a single node's archives with
~200 metrics collected at 30 sec intervals (3 GB), roughly 1 file per
day, ~30 files. Initial load is very quick, 1-2 seconds. When editing
the metrics it takes a while for the field to populate. So for instance,
edit the "1 minute load average" and click on "kernel", I get a blinking
cursor for 10-20 seconds until the list displays. Not unexpected for a
large dataset, but it would be great to get some feedback that something
is happening. Clicking the pencil icon and modifying the metric name
directly gets almost instantaneous results.
3.
I am starting to do my own testing, but has anybody done scalability
studies? My largest dimension is going to be number of hosts. So is it
reasonable that I could plot ~5 metrics over a day but across 100 host
archives? 1000 host archives? With or without the summarization
discussed in 1 above?
Thanks for any insight.
Martins
|