|dim_STAT User's Guide. by Dimitri|
Single-Host Analyzing is very similar to Multi-Host, but gives a wider variety of parameters as it is working only with one particular STAT collect. Let's use as an example the Demo collect, which is provided with the dim_STAT database and let's analyze IOSTAT data. Open your browser and follow step by step how we're connecting to the dim_STAT server.
|Choose Collect and STAT|
|Example IOSTAT: Choose Disks criteria|
Your choice of options is much broader in Single-Host mode. You can analyze your collected data in fine detail, adapting them to your needs... Disks - several possible combinations, but quite similar to other multi-line STATs
Interval is similar to Multi-Host analysis. To simplify, let's look at the last 100 measured data per disk (there are only a few). Values Special Operations - You can analyze on a per disk basis, or SUM/AVG all of them, or group values by the first N characters of the disk name (very useful if you want to analyze I/O activity per controller), or when N is a negative number by the last N characters.
- nothing selected means using all data without refining your select
- you may refine your criteria by selecting only certain disk(s)
- you may exclude your selected disk(s) by clicking the 'Inversed Selection' checkbox
- you may use value-oriented selection (ex. Top-10 Busy% disks)
- you may exclude disks with unwanted data values
- or finally, give a select pattern (very useful if you want to avoid SDS metadevices, etc.)
|Example IOSTAT: Choose STAT Variables|
The data can be presented in three different forms:
Fine, here I want to see:
- Graphics - graphical representation (as we saw already before)
- Table of Results - the raw data is presented as HTML or Text output (table format) and printed on screen or into a temporary file
- Top-N values - in a few clicks check the MAX/MIN values of any STAT variables during the given time period. For example: if there were no disks busier than 30%, you even don't need to look at graphs, or if there are any, you know at once the time slices you need to analyze for a possible jump in activity.
A Bookmark Links may be inserted at the bottom of every viewed graph. Clicking on one of these links will show you another statistics view for exactly the same time period. Click "Start" !!
- with disk Busy%
- and Bookmark Links
|Example IOSTAT: Result Graph|
Some new things here. Under the graph you'll see a list of Bookmark links. If you click on "CPU" (for ex.), a new graph will appear with the CPU activity during the same time period you're observing now. This is useful, because even 3 days later will still point to the same time slice. You'll also find an "Add LOG-Massage" field, the same as with Multi-Host. And a new one: Save Graph as Bookmark.
|Save Graph as Bookmark...|
This is a really cool feature that will save you time. Right now, you can simply give short and long names for your graph view and save it as a new "Bookmark". Once this is done, all the options you selected will be saved (booked) under the name given. And instead of having to click again on all those checkboxes, to get similar data but for another time period or another STAT collect, all you will need to do is just click on the one button with your "Bookmark Name"! NOTE : Since v.9.0 there is a possibility to create Bookmarks for Multi-Host Analyzing too! And all Multi-Host stats are Bookmarks since then ;-) -- To be able to create a "Multi-Host Bookmark" just keep in mind that when you're comparing several hosts you cannot bring on the same graph more than one statistic value on the same time! (for ex: you cannot see both Sys% and Usr% CPU usage on the same time without creating a mess in the graph legend, while you're using only Sys% or Usr% you'll need to show only host names within a legend) - so as far as you're generating a graph with a single statistic value and using only generic data filter conditions in the Bookmark form under your graph the choice will be automatically extended by "Multi-Host" option within a select box! There is a huge benefit to use Bookmarks when you're analyzing many hosts on the same time and on the same graph, for ex:
As well, don't forget to share with others if you're creating new Bookmarks ;-))
- you can follow a fixed list of disk controllers on all servers rather to see a sum of all disks..
- you can follow CPU usage by selected users/processes on all servers rather a whole CPU usage..
- and many others ;-)