Penjelasan Pengujian Beban, Tekanan, dan Ketahanan

Performance testing is an umbrella term. Different test types answer different questions about how a system behaves under demand. Knowing which to run saves time and exposes the right risks.

Load testing measures behavior under an expected level of concurrent users or requests. It answers: does the system meet its SLAs at normal peak traffic?

During load tests you watch:

• Throughput (requests per second)
• Latency percentiles (p50, p95, p99)
• Error rate

Stress testing pushes load beyond normal limits to find the breaking point and observe how the system fails and recovers.

A good stress test reveals whether the system degrades gracefully, such as queuing or rejecting cleanly, rather than crashing or corrupting data.

Soak (endurance) testing runs a sustained moderate load for hours or days to expose slow problems like memory leaks and resource exhaustion.

Spike testing applies sudden sharp jumps in load to check how the system handles abrupt traffic surges, like a flash sale.

Tools describe load as a ramp profile. This pseudo-config ramps to 200 users over a minute, holds, then ramps down.

Always capture a baseline run first. Performance results are meaningless without something to compare against across releases.

Run performance tests in an environment that mirrors production. Testing on under-provisioned hardware gives misleading numbers.

Verify SLAs with load tests, find limits with stress tests, catch leaks with soak tests, and test surges with spike tests. Each covers a distinct risk.

You learned the performance test types:

• Load: behavior at expected peak
• Stress: behavior past the breaking point
• Soak: long-duration stability and leaks
• Spike: sudden traffic surges