how can i measure the data volume that counters or templates produces in the Orion-database?

Hi,

i need to know how high is the data-size, that a template (and better the single counters) produces for every single server, that is under monitoring.

I need this for decisions what kind of template i use, or for understanding if it is worth to switch off the monitoring of some counter in a template to reduce the amount of network traffic and the space occupyed in the orion - Database.

For example:

storage_size <template-name>, <servername>, <storing the data for xxx days>

And the result is the amount of space occupyed in the Database in Mbytes

Thanks in advance

Harald

Find more posts tagged with

template

amount

calculating

produces

data

that

Accepted answers

aLTeReGo

We've done some preliminary tests in the lab. On average, each monitored SQL instance consumed 261MB of storage space in the Orion database with 90 days of historical data was collected. Network traffic for a monitored SQL Instance running on an SNMP managed node averaged approximately 11kbps. Actual millage may vary as there are simply too many variables to provide definitive numbers that would apply globally across all SQL Servers everywhere, but these numbers should provide you with a good rough idea of what you should be looking at.

All comments

aLTeReGo

I'm not entirely sure what problem you are trying to solve, but unfortunately I can't think of any easy way to for you to report on database space/size used by individual component monitor or assigned application. This is especially difficult as not all component monitors are created equal, since some component monitors collect more information than others. Not to mention data retention settings and polling intervals that play a significant role in the amount of data that will be stored in the database. I can tell you that on average, each poll of a component monitor stores roughly 264 bytes to the database.

As for network bandwidth, there are close to 50 different component monitor types in SAM, each varying in the amount of bandwidth they consume. However if we take the one of the lowest bandwidth consuming components (SNMP) which uses approximately 0.01Kbps of bandwidth, and compare that to the highest bandwidth consuming component monitor type (RPC) which uses 18Kbps, you can formulate your own weighted average.

hp_

Hello aLTeReGo,

i try to explain for my concrete case:

I wanna use "AppInsight for SQL". I have ca. 20 central and important MSSQL-Servers and ca. 150 less important SQL - Servers in branch offices (over WAN). I read, that AppInsight for SQL must be used entirely, it is not possible to switch off some counter that i don't need.

I must now decide if i use AppInsight for SQL only for the 20 important SQL - Servers or if i can use it for all the SQL - Servers (20 + 150). And what impact has this to the network traffic and for the storage volume in the Orion-Database.

I don't like to take decissions without knowing the amount of impact that a decision has.

Kind regards

Harald

aLTeReGo

AppInsight for SQL is an especially difficult beast, as there are too many variables to provide a concrete answer. Just some of the variables include the number of databases on each SQL server, the average number of SQL Error Log Entries per-hour, etc. I can say that managing the host via SNMP instead of WMI will lower/reduce the amount of network traffic. At that point all information is gathered using SNMP and SQL queries. Both of which are very bandwidth efficient and latency friendly.

hp_

Thanks, i understand now, that it is difficult to precalculating the amount of data.

I had monitored now already for some months an instance with "Appinsight for SQL".

It would be too a help if i can know how many data has occupyed this monitoring for having an idea how "heavy" is Appinsight for SQL...

How can i measure that?

Kind regards

Harald