Quantcast
Channel: THWACK: Message List
Viewing all articles
Browse latest Browse all 20686

Re: STM Beta - some screenshots

$
0
0

Hey guys,

 

The new beta installed fine for us and has been collecting for a few days without a problem.  I do notice a couple of issues i wanted to paste here.

 

1.  On the new Storage performance charts, clicking the calendar to change the dates when choosing "Between" for time range makes the calendar move under the left NAV (check screenshot below).

2.  Same thing occurs on the "Performance" tab if I do Array Perf and choose between times.  That popped under calendar can't be moved by me.  Funny, if i click the left calendar it pops under, if i click the right calendar it pops a few pixels over.  Check the overlap below.

  • Can you confirm the browser you are using? We believe that this is only a problem on IE8, IE9, and Firefox, but should be OK on IE10 and Chrome. It would be good to have that validated externally.

 

3.  The new charts look great.  I may have to retain data longer and more granular now.


 

4. What does the "RAW" data button do these charts.  All my peaks are a few hundred of IOPS higher using "Raw" data.  Is it Peak data?

  • "Raw" forces the performance chart to load the raw data points. STM has several classifications of data samples: 1) Raw - what was originally collected and 2) Rollup Summary data in Hourly, Daily, Weekly and Monthly intervals. If you are looking at a year's worth of data, loading Raw data in 15 minute samples is not going to result in a good user experience, so we will present the monthly data points. There are similar heuristics at other time range intervals to show you the most granular data that makes sense from a product performance perspective. However, there are definitely times where the roll-up averages will mask potential spikes in the data. That's what the Raw button is there for now, to load the raw data. So potential use case is that you are looking at 8 days worth of data, trying to troubleshoot an issue that happens on Monday's. In this case, we will show you hourly data. Well, if your IOPS were only peaking for 5,10,15 minutes, that spike might be averaged out in the hourly roll-up, so this should help you get to root cause. In general, proceed with caution, as this can load a lot of data from the database.

 

5.  Disk perf bar under the "servers" tab displays nothing, same as #1 above.  That may be by design but looks confusing.

  • Can you post a screenshot here? I'm not sure what's going on. This could be a mapping issue or something else entirely.

 

6. Fast cache data charts is great, I will definitely have to update my collection polling.

 

Looks really good to me, i'll be upgrading it fairly quickly once you guys GA it.



Viewing all articles
Browse latest Browse all 20686

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>