Can the task engine handle millions of completed Tasks

, ,
1

Hi All

I have a query.
What happens to the task engine if the tasks are not deleted from webMethods? That is there are millions of completed tasks in the task inbox. Can webMethods easily handle this? I have a situation where they want to be able to query tasks that are over 2 years old, they want to be able to search the the task inbox.
I have looked at webMethods task user guide and cannot find anything useful.

Regards
Kevin Wong

2

having millons of task instance is not a problem as long as your task query is not running too long and you are not having complex queries on task business data. If you are having custom task search based on business data, having millions of records would impact the performance. Custom business data are not indexed on database. Infact they are stored as serialized object on DB and MWS load the business data from DB and then perform the custom search.

Hope this helps.

Jiten

3

I think that your database size is one of the main issues here. The TaskData is stored in the database as a serialized object. In one of our process the task data object in the db is 3kB large (and this is not a very large one). So millions of tasks result in a large database. You might want to have a look at the table T_TASK_DATA which contains the business data of your tasks.

Also when issuing task queries that search tasks by custom task data is not possible without loading each task from the database (of the respective type). So you should avoid such queries when you expect a large result.

I also experienced that the pub.task.taskclient:searchTasks service is very slow if you query several hundreds of tasks and you want to load the task data (query with includeTaskData=true).