In this short demo I am showing you that Pytrend now can process HPE 3PAR stats dumps and build various system level graphs, such as:
- IOPS over time (read/write/total)
- MB/sec over time (read/write/total)
- Latency over time (read/write/average)
- IO Size over time (read/write/average)
- CPU Utilization over time (system/user)
All above stats are directly parsed from dump into the graphs.
Though, there is another one, which I’ve added myself, because it is very useful to see how latency depends on IOPS, to see whether system resources (disks or controllers)are close to saturation.
Once upon a time, we were using in Mitrend, and I liked it very much. So why wouldn’t Pytrend have it as well? So, there is 6th type of graphs — «Latency on IOPS» as a Pytrend bonus to standard HPE data. It was little bit tricky to implement, I should say, because sorting two dependent lists of data is not the standard thing in Python 🙂
You may also notice, that there are averages and 95-percentile for each value on each graph. They are calculated before plotting, and they do not change when you zoom into region of interest. To help you to get more information about regions of interest (such as peak periods of workload), I have implemented -t1 and -t2 options, which allow you to build plots for shorter intervals of time. Both parameters are optional. You may use one of them, or both, or not use them at all. Value of these options is that you will have your averages and 95-percentiles calculated for specified region of time.
It is very nice, though, there is a small problem I am still planning to solve — various systems use various date formats, so Pytrend may incorrectly compare dates.
There is no problem using t1 and t2 within single calendar day period though.
It is not difficult to fix, but I need time and data from different systems, to test how it works.
Enjoy the demo.
If you want to process data from 3PAR storage — just write me an email: firstname.lastname@example.org