How should you test the analytics platform’s resilience to increased mobile network latency?

Deploying failure injection software directly into the analytics platform allows for controlled and targeted testing of latency. This method precisely simulates increased mobile network latency by adding delays to the analytics traffic as it's processed, directly assessing the platform's resilience. Building a test client on a Compute Engine VM is less effective because it simulates client-side latency, not necessarily how the analytics platform handles increased latency in incoming data. Adding random delay before processing analytics files tests the processing pipeline's tolerance for delayed data arrival, but not the platform's behavior under network latency during data transmission. Creating an opt-in beta tests real-world performance but lacks the controlled environment needed for specific resilience testing and may not reliably generate the desired latency conditions.