Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Document count_* metrics #714

Open
greatvovan opened this issue Nov 5, 2020 · 2 comments
Open

Document count_* metrics #714

greatvovan opened this issue Nov 5, 2020 · 2 comments

Comments

@greatvovan
Copy link

greatvovan commented Nov 5, 2020

In my Grapfana+Graphite+statsd installation I see in the UI of time series of a timer:

image

This may be confusing for newcomers.

  1. What is count_ps? Only in some discussions and issues I found that it is "count-per-second". In the documentation there is no a word about it.

  2. What is the meaning and use of count_90 and count_99? I also suggest to add it to documentation.

@BlueHatbRit
Copy link
Member

Hi @greatvovan, you're right we're lacking some documentation around these, they're being worked on but likewise if you'd like to contribute then that would be awesome!

count_ps is indeed count-per-second which is calculated in addition to your flush interval. Count_90 and Count_99 refer to the 90th and 99th percentile, which we also give with a few of the other metric types as well. It's a great way to understand what's happening often vs rarely in a system.

If you're not familiar with 90th, 99th percentiles in metrics then there's a good explanation here - https://stackoverflow.com/questions/38781499/what-does-90th-95th-99th-pct-matrices-means-in-dashboard-report-of-jmeter.

Hopefully I'll get some time to do a documentation pass over the next few months, things are quite busy for me at the moment though and I seem to be the primary maintainer these days.

@greatvovan
Copy link
Author

I absolutely can do that, but first of all I need to understand it myself.

@BlueHatbRit I know what are percentiles, but I have no idea how to apply this concept to counters. Say I have 10 values:

4 3 5 20 3 3 4 5 30 4

Percentiles consider the values and 90 percentile will cut the highest two (20 and 30). But as for counters, it will work regardless of values, and the result will be always 8. It looks like count_90 is always equal to count * 0.9 and count_99 == count * 0.99.

If the above is correct, I cannot even imagine what can be the use of such metrics.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants