I was reading the performance warehouse best practices here and was wondering about the following section:
"You must NOT connect more than one AppMon Servers to the same (named) database on the same DB server instance! This will wreak havoc with your data as this is not intended."
That being said, can you connect multiple AppMon Servers to the same named database, but have different schemas on that database?
Solved! Go to Solution.
I don't think this would work and even if it would I wouldn't recommend it as the documentation is so insistent that this is not supported. I'm not even sure how you would go about telling the server to point to a particular schema since you only configure the database in the performance warehouse settings, there's no option to point to a particular schema. I think the server requires a dedicated database - is there any reason you can't just have a second database on on the server you are planning on using as the performance warehouse? I think that would be fine.
For Oracle databases, the schema is the username (with the ability to create users who don't have their own schema). So if you connect to a the same oracle database, but with a different user that will give you a new schema.
The DBA group is telling us that the single database with multiple schemas is the better way to go than 2 databases due to resource allocation and other things. I was wondering on the technical details of how the performance warehouse accesses the database, and if it contains itself to the schema, or will go outside of it and what potential problems it can cause.
the wording is a bit tricky to get right across several DBMSes, but what we mean is essentially what Oracle calls a "schema" (which there is the same as a "user"). In some other dbmses the concept is called a database.
Your scenario (two named schemas in one Oracle database) will work, as you can differentiate this with a different user. Note that dynatrace is an application that writes relatively much (compared to other applications), so putting two high load dt servers on the same db server will require very good i/o performance.
Maybe we should re-word the paragraph in the documentation. "Schema" is probably the more universally understood term for this concept. Best regards,