The client is a smart home technology provider serving customers across the United States and Canada. Its services include home security, energy management, home automation, local cloud storage, and high-speed internet.
The challenge
The client asked SoftServe to build a performance testing framework that would serve beyond this one release. It needed to:
- Expand easily for future performance testing and per-release scalability testing
- Show how the wider system behaves when individual systems or components become unavailable
- Identify resource bottlenecks before they affect users
The framework also had to be easy to maintain. The existing team needed to run it with its current infrastructure, tools, and skills.
What SoftServe did
SoftServe's engineers tested the system's main components:
- .NET application hosted in an on-premises data center
- MSSQL database hosted in an on-premises data center
- Salesforce service
- Payment services
- Loan services
- iOS applications
The team used three tools the client already had:
- Splunk: Test logs flow in through a log collector, where results are aggregated and graphed.
- SolarWinds: Infrastructure resource metrics.
- AppDynamics: Detailed application metrics and root-cause investigation.
The previous release had performance problems, so finding them wasn't enough. The team also had to leave time to fix them. Testing ran from March to May, ahead of the June-July release. It found the causes of the poor performance while the client's QA team still had time to remediate them.
SoftServe built the testing approach on tools the client already owned, using them in new ways and adding little new software or hardware. To cut the effort of writing performance tests, the team reused existing functional tests and converted them with minimal modification.
What changed
Performance testing finished on time. The client fixed the issues it found and released the application on schedule. No new tools were introduced, so the client's test automation team took ownership of the framework and now uses it for later projects.
The client's QA Manager put it this way:
The performance testing team at SoftServe helped us deliver on a critical project in a highly-compressed timeframe and in a manner that will allow us to build on the delivered product for future projects. I would absolutely work with this team in the future if I was looking for quick and quality delivery of performance test automation. Thank you!
The release that followed a problem release went out on time. The next performance test now starts from a framework the client's own team already knows.
Let's talk about finding performance problems before your release date does.
