Is there a way to extend the period
of 10 days that CAS keeps datain the5
minutes resolution to 20 days?
If the answer to the above question is yes, what will be the impact on the DB in
means of size and performance?
Thanks in advance for your answers
Solved! Go to Solution.
Yes you can change the 10 days by going to <casip>/DiagConsole#/sp (assuming you are on 12.3) and updating the property AMD_STORAGE_PERIOD
As you say this will have a large effect on your database size. The database size increase is difficult to say for certain as it depends on things like traffic profiles but you can estimate it by taking the size of all your rtm*stats tables and multiplying it by the increase you intend to make. In your case doubling it.
In terms of performance this really relates to the specs of your db server but as a general rule of thumb I try to keep CAS DBs below 300GB.
Hope this helps.
Yes it is possible to extend that.
When you go to http://cas_address/atscon
Then click on the "storage period"
AMD_STORAGE_PERIOD is what you are looking for.
When it comes to the impact, then:
a) for sure the DB will grow by some percent. I am not able to tell by how much, as the 5 mins resolution tables are just a fraction of the overall number of tables in the database.
b) as you would like to keep more data, then you will probably want to use that in your reports (so instead of showing 10 days on reports, you will show 20 days on the reports). In that case you may experience that 20 days reports will load slower than 10 days reports
c) there is some part of data, that will be removed when inactive for last 10 days. If you change the storage to 20, then those will be kept for 20 days instead. So if you have frequently changing operation names, then those old ones will be there for 20 days instead of 10. In essence - if you change the storage period, please observe the behavior of 'server cache' number on the CAS System Status Report (tools -> diagnostics -> system status)
d) if the 'server cache' from point 'c', goes significantly up, then the data processing time on the CAS, may raise (as logically there will be lot more entries to go through, in order to figure out if some item has to be updated, or new one has to be inserted)
So all of the above, will occur on you environment, but some may be move visible than other. It all depends on what is being monitored by the AMD.
If you wish to introduce the change you mentioned, just observe the system for next 10 days after the change, in order to evaluate pros and cons. People tend to tailor that according to their own needs.
You might just win the prize for the biggest DB 🙂
So the first question you need to work out is does my DB need to be that big. And by that i mean is there a mistake in the config, i.e. keeping data too long, too granular a level, unique parameters or is it all required.
I would start by checking the table sizes in the DB by going to http:///LSServlet?dmiAction=Generate&lsActi...>/LSServlet?dmiAction=Generate&lsAction=LoadByName&lsEntryName=Database%20Status, then the aggregation setting for my users, then my session breakdown.
If the answer is yes you need all that data then your solution is not sized appropriately and you should look into the Load Balanced Clustering feature that was introduced for CAS servers in 12.3 (https://community.dynatrace.com/community/display/...
If you need further assistance with this the Expert Services team offer a DC-RUM health check which includes short and long term recommendations to get the most out of your DC-RUM deployment. I would like further information on this please reach out to your CSM.