Continuous load and performance testing results using k6
We are committed to providing reliable and high-performance services. This page details the results of our continuous load and performance testing, conducted using the open-source tool k6 by Grafana Labs. Our tests simulate real user traffic to validate stability and response times under high load conditions.
The hardware and caching features allocated to your hosting environment directly impact performance under load. Our k6 testing reflects the difference between these tiers.
The data below represents the combined average performance across all tested scenarios for both plan types, simulating a total of 1,500 Virtual Users (VUs).
These results illustrate how the Business Plan's increased CPU, RAM, and especially the Redis Object Cache, significantly reduce latency and increase throughput for key dynamic operations.
| Scenario | Simulated VUs | P95 Latency (ms) | Max Throughput (RPS) | Error Rate | |||
|---|---|---|---|---|---|---|---|
| Personal | Business | Personal | Business | Personal | Business | ||
| Static Homepage Load (Simple GET) | 1,000 | 45.8 | 42.1 | 100 | 105 | 0.00% | 0.00% |
| Dynamic Blog List Page (Low Cache Hit) | 500 | 185.3 | 75.2 | 40 | 90 | 0.01% | 0.00% |
| User Login/Auth Cycle (Database Check) | 250 | 220.5 | 110.8 | 25 | 45 | 0.15% | 0.00% |
| Complex Checkout Process (Multi-step POST) | 250 | 310.9 | 210.3 | 15 | 20 | 1.20% | 0.01% |
Key Takeaway: The Business Plan demonstrates significantly lower latency (faster response times) for dynamic and database-intensive tasks (Blog List, Login), due to the combined power of 2x Resources and Redis Object Caching. While the Checkout process remains complex and slower on both tiers, the Business Plan handles the load with far fewer errors and lower latency.
We continuously monitor the performance of both plans.
Recent Trend: The Business Plan's performance remains highly stable, typically maintaining P95 latencies below 120ms for dynamic content. The Personal Plan shows higher volatility and slower performance during peak usage.
