Case Engine Release 1.1.42
Storage & performance
This release contains major changes to the way the case engine is storing information and processing it internally. This leads too great savings on the storage side, as well as considerable performance improvements.
Savings on your existing database can be achieved as well.
A tool is available that can do this for you, even on a live production system.
Contact us for more information about the database conversion tool.
Dependency updates
Updates that fix security issues in the underlying dependencies of the case engine have been applied in this release.
This includes database driver versions, Apache Pekko libraries, authentication libraries and more. All have been updated.
Less events
When something happens in the case plan, e.g. a task becoming active, it would to many events on that same task.
In this release, a major change has been applied on the case plan events, ensuring that no more than one event per task (or stage, milestone, timer, etc.) is generated.
Also the names of the events are much more descriptive.
As an example, a task becoming active would generate an event PlanItemTransitioned. This has become straightforward TaskActivated, which is more functional and to the point.
Less storage
Perhaps this should read "way less storage". In the classic events, the XML based information containing the case definition was printed again and again in events. For large case models, this can make a database grow fast.
In this release, the unique case definitions are stored with a special hash in a database table, and the events only carry the hash. This leads to massive storage savings.
Better performance
Case Engine projections that run in-memory are now updated directly after the events have been persisted. This leads to improved response times on queries that are ran after the events got persisted.
This option is enabled by default.
Release 1.1.39 introduced an additional tag on every event containing root actor id. The purpose of which is to be able to retrieve all events of a case chain in one shot.
Customers reported a negative performance impact with this, and from this release onwards that functionality is dropped.
Housekeeping, improvements & fixes
Various small improvements and fixes have made it to the release.
Some of the improvements:
- Major refactoring happened "under the hood" in the sentry network.
This enabled fixing issues related to Case File Items of type array. - Duration parsing has extended
- Every task that has a due date set now also carries a duration, based on the due date of that task.
- It is now possible to give both a fixed date & time as well as a duration to timers, and also to task due dates.
- More migration scenarios are now supported.
- Low level log messages have been improved on clarity and readability.