Looking into transaction flows, I notice that there are 2 database assets, one recognized as oracle, one just "database". Looking into the details, I can find that both are about the same oracle database. This is quite confusing.
A more experienced collegue of mine, told me that this was reported earlier, but should already be solved in 6.1
What am I missing
If this is happening it means that our JDBC/ADO.NET instrumentation doesnt correctly pick up the Connection Type. This could be caused by an instrumentation problem or a problem with parsing the connection string. Is there anything different in the connection string if you compare statements that are executed one or the other database?
Nevertheless you should probably open a support ticket and see if this might be a known issue already
Sorry to have waited so long to reply. Today only I had the oppertunity to sit together with java-specialists and came to the conclusion that the entries we find for the "undefined" database are probably queries to JPA-entities that are stored on a in-memory database.
Does this make any sence ? I will further discuss with the java team on how to interprete this and whether we need additional configuration (possibilities) to make more visible that this is about JPA (Java Persistency Layer)
When JPA is used (java persistency Api), it can be configured to do some caching. In case this happens, DT wil visualise this as an extra instance of database-type in purepaths, workflows etc.
Untill now, we have to live with it...For me it was most important to be able to identify what is happening. Why do we have that extra database visualised.
Maybe we should create an RFE for this, as it looks as we are not the only ones facing this issue.
When there is some (prepared) statement caching active, AppMon cannot detect the JDBC URL, as there are no calls to getConnection() anymore. This is a known limitation of the product.