if any guide or doc, that would really help...
Solarwinds, by default monitors the stats using OIDs from the manufacturer.
When you say calculations, how do you mean?
Performance data is collected 'as is' from the device, if that data is performance related (CPU, memory, disk, etc) it is usually stored in the database raw and then calculated into min/avg/max values, if not in the short term then certainly in the aggregated tables. Data aggregation is configured on your Orion Polling Settings (<Orion address>/Orion/Admin/PollingSettings.aspx), which is a good place to start.
Raw data is rolled into hourly after X days (check your Polling Settings, usually 7), then hourly data into daily data after X days (we have 31 days, though I think it's 30 by default) and then stored for a year (check your settings) before by dropped from the database. The long term data (up to 180 days worth) is used for trending and capacity planning calculations. Check Monitor capacity usage trends on the network and forecast capacity issues - SolarWinds Worldwide, LLC. Help and Support for more details.
Just remember the more data you keep, the bigger the database gets, and when things get too big performance suffers, check <Orion address>/Orion/Admin/Details/DatabaseStats.aspx to get an idea of your table sizes.
Hope it helps, if not stick in a few more details of what you're after.
I was mainly referring to the formula that runs behind. So if i understand correctly, it will be specific table where values gets stored and the max/avg etc gets calculated? OR there is something more to it?
The aggregation and polling part i am aware... Its more like if any of my clients ask me, i should be able to provide them some input.
1 of 1 people found this helpful
Yes, log onto your Orion server and open up the Database Manager Tool (C:\Program Files (x86)\SolarWinds\Orion\DatabaseManager.exe) and click on "Add default server".
This will attach to your SolarWinds database and allow you to explore the tables.
Firstly Nodes is the place to start, with the all important NodeID, this little bit of information ties most of the database together.
Next is the area of interest, let's take CPU, scroll on down to the CPULoad (View), right click and Select "Query Table", once the SQL has loaded in the left pane, click Execute Query and it will return the top 1000 rows of that view (a view being a view of a collection of connected tables, usually providing an easy way to view the data). This view is made up CPULoad_Detail, CPULoad_Hourly and CPULoad_Daily tables, all tied together with the NodeID column.
Here you will be able see how the data is stored, with the min,max and avg values, the data within the last 7 days will be in it's native polling timeslots (300 second poll, 1 line of data every for each poll), past 7 days old these 5 minutes are averaged over the hour and the average, min and max values recorded, past 30 days it's aggregated into daily.
Check out the Interface and Volume tables, they are quite similar (min,max,avg wise anyway).
The aggregation is done by stored procedures, within the database, typically triggered overnight to run the aggregations, when there is less user activity.
If you use SQL Server Management Studio to connect to your database you can view these.
Have a dig around, it should help.
Just remember not to hack around or change anything, without taking backups!