We have a single agent (v9.6 CU23) using MySQL. On it there are several elements which create DVE's. If 1 parent-element creates 100 DVE's, then all trend data will be put in the DATA/DATAAVG table of the parent. Causing this table to grow and eventually get into trouble due to its size (slow data ingest/update/..).
Is there a way to mitigate this? More beyond the point of looking into the trend template(s).
Is there a way to generate a DATA(AVG) table per DVE in order to avoid such large parent DATA/DATAAVG tables?
(The discussions to move to Cassandra are on-going but complex)
It's not possible to move the data to child element, as that would require too many changes in dataminer.
What could be an option to look at is partitionedTrending (databaseOptions attribute | DataMiner Docs)
The major slowdown on MySQL is mainly the cleaning. And all the other things just slow down during that cleaning. partitioning could remedy that.
Note: going to Cassandra would obviously still be the preferred way 🙂
Robin, we have been having these same issues for years on multiple systems and like you, moving to Cassandra is not an easy move. If you would be interested, we could connect offline and share any experiences and possible solutions.