Hello, I am new here and new to Tableau.
I have 2 situations that I am unsure of if its normal behavior and how to stop/prevent them from occurring in the future.
1) We have a live connection to an oracle database that is hosted by one of our vendors, the vendor admin is reporting that they are seeing long running connections hanging out there for a day or more. They are concerned that this can cause performance problems with the database. I found a kb entry that explained that live connections were kept open in cache until it is expired and that space is needed for other work, it suggested making a change to the vizqlserver.protocolcachelifetime_in_minutes and vizsqlserver.protocolcachesize parameters in the config to force the connection to cache out. I did make those changes and have not heard from my vendor that they have seen any new long running connections since then so we will see if that did the trick. Is there anything else I could look into to prevent this behavior and is this explanation of what is happening sound correct? We verified that nobody was accessing the workbook through the night or in the morning the last 2 times it was reported so there is no way that the cache lifetime timer was getting reset in these instances but we are new on tableau server and only have a couple of published workbooks and data sources at this time so it is possible that the cache was just not needed for something else. This connection is embedded in the workbook also, not published separately.
2) We have another live connection, this one to mssql db, and it seems that this one is querying the database just about every hour regardless if anyone has opened any of the views or dashboards using the connection. I can see this happening when I view the logs. Again, is this normal behavior for a published live connection? Is there a way to prevent or stop this and only have the query run once someone accesses a workbook or view associated to the connection? This particular connection was published separate from the workbook and it does not appear to me that this same behavior is occurring with the live connection mentioned above.
we are running server 2018.1 currently.