Don’t test performance in production. Here is a checklist for a systematic approach.
Preparation¶
- ☐ Performance requirements defined (SLA/SLO)
- ☐ Baseline measured
- ☐ Test environment matches production
- ☐ Test data prepared
- ☐ Monitoring active
Test Types¶
- ☐ Load test — normal load
- ☐ Stress test — beyond the limit
- ☐ Spike test — sudden surge
- ☐ Soak/endurance test — long-term load
- ☐ Scalability test — with increasing load
What to Measure¶
- ☐ Response time (P50, P95, P99)
- ☐ Throughput (req/s)
- ☐ Error rate
- ☐ Resource utilization (CPU, memory, I/O, network)
- ☐ Concurrent users
After Testing¶
- ☐ Results documented
- ☐ Bottlenecks identified
- ☐ Optimizations implemented
- ☐ Re-test after optimization
- ☐ Results shared with the team
Tools¶
k6, Locust, Gatling, JMeter, hey, wrk. Choose based on your language and needs.
performancetestingqa