Timeouts are essential for performance because they help to prevent system resources from being tied up by long-running or unresponsive processes.
In a distributed system, for example, when one component sends a request to another, it expects a response within a reasonable amount of time. If the reply takes too long, the requesting component may become unresponsive or tie up system resources, leading to performance issues and potentially even system failure.
Timeouts provide a mechanism for limiting the time a process can wait for a response and terminating the process if the response does not arrive within the allotted time. By doing so, timeouts can help to ensure that unresponsive processes do not tie up system resources and that the system remains performant and responsive.
The most common timeout issues in a T24 Transact core banking environment are related to Enquiry.
Suppose the jbase agent processing the T24 Browser request takes more time because of a few reasons like massive enquiry execution, big record processing, and bulk commits than the time specified in the parameter 'Connection timeout' in the file browserParameters.xml. In that case, the T24 browser front-end throws an error message "Connection timeout or invalid message type returned from JMS queue." To avoid the above T24 browser crash, please do the following.
Increase the 'connection timeout' parameter's value inside the browser. Parameters.xml file
Increase the maximum connection count from 10 to 20 to provide an additional buffer during peak hours.
Keep the 'action timeout' parameter's value not more than 180 secs to kill the jbase agents which are processing huge requests/commands for more than 3mins
Use Enquiry reports if the output data load is heavy.
OFS Response Timeout Problem
These type of timeouts are similar to Browser timeout, but it occurs from 3rd party system while trying to fetch details from T24. So we need to check if all the functionally related changes are in place and test the enquiry by launching it in the browser.
File Timeout Problem
Avoid using the Nofile routine if it's using only two applications in an enquiry
The build routine should follow all the programming guidelines mentioned in our technology changes.
Sort out selection criteria in the enquiry itself.
NoFile Timeout Problem
Follow all the programming guidelines mentioned in our technical changes to fine-tune the routine.
Use the Filter Routine to filter out the records if required.
Avoid using more conversion routines and include the logic in the No File routine itself.
Timeouts can help improve a system's reliability by detecting and recovering from faults or errors. For example, suppose a component fails to respond to a request within the allotted time. A timeout can trigger a failover mechanism that switches to a backup component or retries the request with a different configuration. This can help minimize downtime and improve the system's overall reliability.
Contact us anytime regarding T24 performance-related topics. We are happy to help.
Keep up the great work! Happy Performance Engineering!