República AFAP is the leading private organization for administering pension savings funds in Uruguay.
Workers must be registered with the Social Security Institute and also choose an entity with which to set up a complementary individual savings system.
República AFAP is the most trusted provider in the country for this financial service, backed by the government.
República AFAP was going through a migration of its CRM platform, from SAP to a custom CRM implemented by a third party development company, DVelop Software.
The migration was of such a high impact to its business operations that it was essential to ensure that the application would be able to handle the expected load without issues.
Some of the project’s stakeholders were resistant to change, so when it came to the new platform, there was no room for complaints.
Abstracta created a plan involving three stages of testing utilizing the following tools for test execution and monitoring:
- Activity Monitor
- Dottrace jetbrains
First, Abstracta performance engineers automated scripts that covered the main functions of the system which had already been released.
The first test execution consisted of five test cases that formed the performance testing scenario.
From those executions, the engineers identified several opportunities for enhancing response times and detected that the connection pool was too small, which caused an error.
In stage two, the Abstracta team added a new, problematic workflow involving email synchronization to the test suite.
Lastly, the team added four new test cases to the test scenario and successfully executed them along with the other test cases prior to go-live.
We made new improvements in response times and optimizations in sql series.
Abstracta also assisted República AFAP during its go live event to make sure that everything ran without any incidents.
Thanks to the help of Abstracta’s performance engineers, República AFAP was able to gain a clear understanding of its new CRM application’s behavior under a concurrency of users.
- Reduced the average system login response times by 5x
- Found and solved several critical errors, i.e. the size of the connection pool was too small, causing an error in the application under load
- Detected the causes of bottlenecks at the database level
- Immediately solved the problems that were detected after the go-live before they could reach end users