Waiting for something for a long time can be very frustrating. As a human, a waiting time of up to 3 seconds is okay. We get frustrated between 3 and 5 seconds, and processing times beyond 8 seconds result in a massive loss of engagement.
We use timeouts in T24 core banking systems to protect critical resources from overloading. As a result, we have max waiting times or timeouts on multiple layers from the BrowserWeb down to EJB, MDB, Messaging, and databases. If you ignore these basic settings or do not harmonize them accordingly, you will see frustrated T24 users and severe performance issues.
Some of the most critical timeout issues are
OFS ERROR TIMEOUT
ACTION TIMEOUT
BROWSER WEB TIMEOUT
OFS ERROR_TIMEOUT reported in global.log
When the processing of the incoming request exceeds the timeout value, the listener will return an OFSERROR_TIMEOUT. This scenario is quite typical. Therefore, we need to increase the value of timeout for these cases.
jbase_agent and action_timeout
You will see an error "Action Timeout occurred" if the jbase_agent was killed after the configured timeout reached. Review the actionTimeout property in the JCA connection factory and adjust it if needed. There might be long-running requests, and if they get killed all the time, such tasks will never be finalized. Therefore, optimize these slow requests and ensure you avoid such action timeout occurred errors.
T24 BrowserWeb Performance
Waiting in front of loading or not responding T24 BrowserWeb screen is a pain. Understandably, reoccurring slowdowns result in frustrated employees and customers. They get blocked during their daily work, and their productivity declines. As a result, they will avoid using these slow T24 systems, causing a massive business problem.
A known issue is the T24 BrowserWeb Connection Timeout error due to slow requests or blocked resources. Usually, it's configured in the BrowserWeb.war, and its parameter value is set to 300 seconds. Therefore, a timeout error is raised if a user request takes more than 300 sec.
You don't need to accept such slow T24 BrowserWeb response times. However, before you start the optimization, you should get a good understanding of the slow business functions.
For instance, online reports are prone to performance problems because they load time-consuming data. However, online reporting can be optimized in certain situations by improving the enquiries or adding indexes to the fields.
Another known issue is the wrong logging configuration. You should not trace all the requests but use the ERROR logging level, which can be configured in the TAFJTrace.properties file.
It would help if you also had a transaction tracing or full-stack monitoring solution to detect slow requests before they impact your end-user or close business operations. Proactive tracing, problem, and root cause detection are critical for such T24 environments.
Further hints to speed up BrowserWeb Performance
Configure the class caching accordingly
Make sure all the TSA. SERVICE's running in AUTO mode are running without any crashes.
Try to avoid generating reports from production, especially during business hours. It will give better performance to the branch users during business hours.
We will start T24 performance tuning workshops for our customers soon. So follow us to receive more updates about this topic.
Keep up the great work! Happy Performance Engineering!