Component to persist rows from a database

[ewonka]

Hi all,

Currently we have registered a service to an exception Event. Whenever an exception occurs the service is fired and performs notification type functions such as paging and email. The paging information is persisted in a database and contains columns with phone numbers and email addresses. We are concerned that a large number of exceptions could wreak havoc on the database and smtp servers.
Is it possible to write a service would read the database “once”, persist and expose the data in order to at least eliminate possible database performance issues? The table is only 77 rows at present so I don’t think it would be too much of a memory hog. Not sure how a component would expose this data?

[Yemi_Bedu]

Hello,

Well the purpose of the database is for high-performance persistence and retrieval of data instead of on a resource like the file-system. If you have performance issues with your service calls to the db, you can make a service to wrap the db call.

That service you should set to state-less with caching turned on but do not pre-fetch. That will allow the data to be in memory on the IS side of call. You can tweak how long to save the cache results at this point. Please remember to use clearPipeline before the call to avoid extra and wrong information from being cached.

Good day.

Yemi Bedu

[ewonka]

Yemi,

Thank you, that sounds very good. I need to work with your solution a bit. I appreciate your response.

[ewonka]

Yemi,

Actually, we have a java component that uses a JDBC connector to perform the database query. I am assuming that i’d set the JDBC connector attributes as you described? Would I also set the component attributes as well?

[Yemi_Bedu]

Hello,

I don’t know if you found out yet. When I said service, I meant a flow service. So you would have the flow service call your java/adapter/whatever service. The first call should be to clearPipeline (with service inputs) so you only have what you want to work with on the pipeline. then you can pass those inputs directly to the call to the inner service (java/adapter/whatever) that does all the real work. The cache settings would be on the flow service, as it is only a wrapper. Match similar names for the inputs and outputs between your flow and inner service. Good day.

Yemi Bedu