openUTM offers you numerous functions which you can use to obtain up-to-date information about the utilization level of the application, to diagnose bottlenecks and to initiate actions to improve performance.
Reasons for performance bottlenecks can include such things as:
Increased requirements on service calls during peak times
Too many users/clients are working with the application at the same time
The processes that are available to the application are occupied by jobs for an extended period because they have to wait for resources locked by other processes
Processing of a large number of asynchronous jobs impairs dialog operation
Too many long-running program units are running at the same time, e.g. program units which conduct a search of all data for specific information
Many program units containing blocking calls are running at the same time, e.g. the KDCS call PGWT or the XATMI call tpcall. During the waiting period, each of these program units occupies a process in the application on an exclusive basis.
With distributed processing using OSI TP or LU6.1, the system waits long for an association or session to be assigned
Frequent I/O accesses to the page pool
Frequent read accesses may indicate that the cache generated for the UTM application is too small.Bottlenecks to connections to communication partners in the application