Who loves slow and not reliable services? I assume that even very patient users get frustrated after a few seconds and hate such crawling applications. Imagine working in a back office and having to process hundreds of very similar tasks in your core banking application. Sadly, some of the masks are loading very slow, and you realize more than 50 percent of your time is entirely wasted by sitting in front of a white screen and waiting for the pages to render or load data from the backend. Such scenarios are incredibly frustrating, lead to demotivated employees, and burn a lot of money.
The good thing is, Core Banking systems such as T24 can be optimized and tuned for speed. This post will give you some ideas on how to get rid of such bottlenecks in T24.
There are always good reasons if services do not fulfill the performance expectations of our end users. Typically, the entire non-functional testing is skipped due to a lack of know-how. The main features are getting tested from a functional point of view, and if they are working as expected, the customized T24 application is deployed on production. Once the team uses it in a production environment, they realize that response times are not acceptable. From an infrastructure perspective, there are no bottlenecks.
To avoid such risks, you should consider these four steps
Performance Requirements
Performance Testing
Performance Monitoring on Test and Production
Continuous Performance Optimization
Create meaningful performance requirements that include the expected number of users or transactions during regular and peak hours. A sample performance requirement could be:
During peak hours we expect 5000 report requests. 95 % of all report response times during such peak hours must be < 2 seconds.
Set up a performance testing pipeline or automated SLO/SLI validation to validate these requirements fully automated as part of every release. This approach brings a log of freedom to your teams.
Also, ensure transaction tracing and application performance monitoring across all layers on all stages from development through production is in place. Create T24 performance monitoring cockpits and chart BrowserWeb, Middleware, Data Layer, and Infrastructure performance metrics. This holistic perspective is the starting point for your tuning activities.
Don't believe that response times will remain the same once T24 is deployed on production.
The data volume grows, the users find creative ways to work with their applications, the noise in environments will create challenges, and soon after deployment, response times often decline. Your only chance to avoid this pitfall is to use a full-stack transaction tracing and performance monitoring of all the layers.
Performetriks provides performance testing, tuning, and monitoring services to many players in the financial industry. Our T24 performance engineering framework is unique. We make performance part of our customers' value stream and ensure that T24 users love their fast responding services.
Keep doing the great work! Happy Performance Engineering!