agency-performance-benchmarker

Define test stages in the k6 options object with duration and virtual user targets, then write a default function that makes HTTP requests to your endpoints. Add thresholds like p(95) response time under 500ms and error rate under 1% to enforce SLAs automatically.

Google defines good Core Web Vitals as Largest Contentful Paint under 2.5 seconds, First Input Delay under 100 milliseconds, and Cumulative Layout Shift under 0.1. These should be measured at the 75th percentile of real user data, not just synthetic tests.

Analyze slow query logs, examine query execution plans, and check connection pool utilization under load. The methodology covers query optimization, indexing strategies, and connection pooling analysis as part of systematic bottleneck identification.

Yes, k6 supports authenticated testing by extracting tokens from login responses and passing them as Authorization headers in subsequent requests. The example workflow logs in first, then tests authenticated endpoints like dashboards using the returned bearer token.

Load testing validates performance under expected normal and peak traffic, while stress testing pushes the system beyond capacity to find breaking points and observe recovery behavior. Both are needed: load tests confirm SLAs, stress tests reveal failure modes.

Performance regression tests should run on every deployment with automated quality gates that block releases failing performance budgets. Full load and endurance tests are better suited for pre-release staging environments due to their longer duration.