Skip to main content

Database performance

After several years or product usage, the database can grow considerably and can affect performance and some important functions of the product. Bravura Security Fabric is constantly updating its backend database with historical data, some of which may be useful for audits, but it doesn't offer any automatic way to shrink data from the database by default.

In addition to hardware improvements, the following maintenance tasks will also help improve performance. Such maintenance tasks must be performed carefully, following your organization's mandatory application uptime and data retention policies.

It is highly recommended that these tasks are performed by a database administrator.

There are four ways to reduce data that Bravura Security Fabric processes and keep performance from degrading over time:

  • Only list and load objects the Bravura Security Fabric solution needs from the target systems; use filters to reduce the data collected to the absolute minimum.

  • Plan ahead if you're going to partition tables in the database, to put older data on slower, cheaper storage, and newer data on faster, more expensive storage.

    See dbarc' s partitioning options below.

  • Periodically "hide" data from the main table views in the database

    See dbarc's hide options below.

    This doesn't reduce the database size, but it speeds up some operations which run on table views rather than on the full tables.

  • Periodically trim the data in the databases.

    See dbarc's delete options below.

    To reduce the amount of unused space, a database "shrink" or "rebuild-indexes" maintenance operation can be performed (best to be done in application low-use times or planned outages, especially not while the auto discovery (psupdate) process is scheduled to run).

All of the above direct database manipulation has to be applied on all databases in an instance separately because the SQL run against the database in an instance does not automatically replicate between application nodes.