I have been playing with Grafana for a while, and got stuck with the following issue.
I need to produce a chart that displays the round trip time for some REST calls my application is doing.
I am using Dropwizard metrics and starting a PausableTimer when I fire the call, and stop it when the expected response comes back. From Grafana, then, I can see the data source I need but when I build the graph, I am forced to apply some statistical calculation on my data (stddev, mean, max, min, etc.) whereas all I need is to chart the numbers from the series.
Below is a screenshot of the config part I am referring to:
The dropdown contains a number of transformations to apply to the priginal data, and I cannot find a way to tell Grafana that I wans to display the data the way it is.
I am not really sure how to do this, had anyone had this issue before -- and solved it?
Thanks in advance.
Related
I am looking for a feature in Apache JMeter. I need the equivalent feature that I use with MicroFocus LoadRunner called "User-Defined Data Point Graphs".
From LR User Guide
The User-Defined Data Point graphs display the values of user-defined data points. You define a
data point in your Vuser script by inserting an lr_user_data_point function at the appropriate place
In pratice I want to use "JSR223 PostProcessor" and via code add values to a time serie. This time serie will be plotted on time based graph. In LR with line of code
lr_user_data_point ("data_point_1",1);
The UI shows a new data serie on plot that can be used as additional report (generate plot, median, table, etc...).
Thanks
Vincenzo
No there is no such feature in JMeter.
You can request an enhancement but ensure you explain:
What is the aim of this feature ?
What benefits it brings
See:
https://jmeter.apache.org/issues.html
Edit on 3 august 2018:
In upcoming JMeter 5.0, it will be possible to graph over time custom variables. See https://bz.apache.org/bugzilla/show_bug.cgi?id=62166
at this point there are three possible solution:
Open a new feature request as suggested by #UBIKLOADPACK
Develop a JMeter plugins using TestBean framework as described into link
Export a time-series to Grafana (or similar) using BackEnd Listener
Regards
I have discovered Kibana last week, I find this tool very useful but.. I can't help but wonder how can I bypass Kibana's limitation to one time range for whole dashboard (https://github.com/elastic/kibana/issues/3578)?
I need to present data for different time ranges on one page. It will include pie charts, data tables etc..
Unfortunately, you can't do that without changing Kibana's code to create a solution of your own.
Here is the github issue tracking the request:
https://github.com/elastic/kibana/issues/3578
I have a Kibana search that returns the load average in a field called metric_value_number. This indexed field has a format of Number, and a Numeral.js format pattern of "0,0.[00]". This is default values.
Here is a screenshot of my search with sample data
When I try to make a visualization with a sum-aggregation, it gets aggregated to zero. This applies to Visualization types of at least Line chart, Area chart and Metric.
How can I properly visualize the system average load with the data I have?
Found this thread, because I had some trouble myself getting our load visualized. However, it turns out I had to type cast to number first by using %{NUMBER:load_1:float} and actually needed an average aggregation.
Had to delete the Logstash index and "reload the field list" in Kibana to get it working.
Tried to reproduce your problem by using your settings. Unfortunately for you, I got it working. I am currently using Kibana Version 4.1.2-snapshot
Build 7500.
This is no longer necessary to keep in forum for lack of poor format nor statistical data keeping.
When the graph says it's running it doesn't really mean any activity is taking place, it just means the graph object told the filters to run and set its "state" variable to "running". However if any filter is stuck for any reason and doesn't send or receive the data everything freezes but the graph state remains "running". Some editors like GraphEditPlus can show states of each particular filter in the graph, so if some of them report to be stopped or paused while some others are still "running" you will see this. In most cases however you'll just see all filters "running" and won't know which of them is stuck unless you insert sample grabbers between them and tell GraphEditPlus to show grabbed samples in real time, this is a good way to understand what's really going on.
I'm using the Web Audio API to make a graphic EQ with four BiQuad filters; a highpass, two bandpass and a low pass:
You can see each node representing each filter's frequency. It's working nicely but I'd like to draw the shape of the filter's roll-off you'd usually find in other graphic EQ's.
I have the Q-factor of each filter and the central frequency. What I would like to do is get a formula that allows me to get the frequency of the roll-off at a specific db value. For example what is the frequency of the roll-off at -200db?
It's been a while since I've done any heavy mathematical stuff and so any help would be greatly appreciated.
Have you looked at the getFrequencyResponse method of BiquadFilterNode? https://dvcs.w3.org/hg/audio/raw-file/tip/webaudio/specification.html#Methods.
This demo from my IO2012 Web Audio talk uses it: http://webaudio-io2012.appspot.com/frames/frequency-response.html. You'll have to calculate the responses of multiple filters stacked together from there, of course.