search

distributed-system-analysis / pbench

A benchmarking and performance analysis framework
http://distributed-system-analysis.github.io/pbench/
GNU General Public License v3.0
188 stars 108 forks source link

low-priority: order of graphs in sar/network_l2.html #371

Open bengland2 opened 8 years ago

bengland2 commented 8 years ago

I love the new graphs, but...

can we make the first graph in sar/network_l2.html be Megabits/sec instead of drops/sec? Here's an example of what I mean. And the 2nd graph should be packets/sec. This is the principle of making the most commonly used graphs appear first. This principle might apply to all of the html generated by pbench.

Another idea that would make the graphs easier to use: have a table-of-contents-like header at the top of the html page that would take you to any of the graphs below with a single click, as well as telling you what graphs are in that page so you don't have to linear search with the scroll bar. Same comment for iozone/disk.html .

Perhaps pbench html doesn't matter anymore if we are using ELKStack to process all data, but it would be nice to have pbench graphs operate well in a stand-alone mode for people who are trying it out for the first time.

ndokos commented 8 years ago

@k-rister: WDYT?

k-rister commented 8 years ago

I like the ideas. That said, this really isn't a decision that is to be made by the chart library itself. The HTML is constructed by the GenData Perl module with the calls to GenData::gen_data coming from the individual tool scripts. It probably wouldn't be too much effort to add the links at the top of the page to the each individual chart. Controlling the order would require adding information to the hash that is passed to GenData::gen_data which a quick little git-grep shows happens from 19 different scripts.

ndokos commented 8 years ago

RIght - sorry. I'll take a look at what it would take to do the reordering and the TOC.

k-rister commented 8 years ago

I'm not opposed to helping with this, but I think some time might need to be spent to determine the best way to implement it. Given that I have only touched GenData.pm to a very limited degree I would like to hear from those who have had a bit more experience with it to see what they think is the best way to do this.