How to Measure TTFT and Inter-Token Latency With Recorded Traffic
Run the mock-lab LLM streaming lab: record streamed chat traffic with proxymock, replay it, and fail the run when TTFT or inter-token latency regresses.
Posts about load testing
Browse 48 posts in this category
Measure LLM Latency on the Prompts Your Users Actually Send
Speedscale replays now measure AIPerf-defined TTFT, inter-token latency and token throughput on your recorded LLM traffic, next to what production measured.
Load Test PostgreSQL Instantly using Production Recordings
Record the queries your app sends to PostgreSQL in Kubernetes, then replay them against a QA database as a load and regression test. No scripts to write.
A Developer's Guide to Continuous Performance Testing
Add continuous performance tests to CI with bounded smoke checks and scheduled load runs. Choose meaningful gates and handle noisy results with baselines.
API Testing Tools Best Practices Guide
Choose API testing tools for exploration, CI regression, and dependency isolation. Compare assertions, protocols, workflows, and migration requirements.
Determine Throughput Performance Testing
Measure throughput with defined transaction boundaries, sustainable load stages, latency and error goals, and enough load-generator capacity.
How to Calculate TPS for Kubernetes Performance
Calculate TPS from completed transactions and a measured time window. Distinguish requests, business outcomes, retries, latency, and errors.
GoReplay vs Speedscale for Kubernetes Load Testing
Compare GoReplay and Speedscale capture, TLS requirements, dependency mocks, and local or Kubernetes replay. Verify the exact workflow before choosing.
How to Test Autoscaling in Kubernetes
Test Kubernetes autoscaling with controlled load. Verify HPA metrics, readiness, scaling policies, and scale-down alongside latency and errors.
Try Speedscale free
Choose the desktop proxymock or the hosted cloud trial to get started.