reliability

name	reliability
description	Load testing & resilience — what breaks under stress and what happens then. Failure-mode matrices, k6/Locust/vegeta load tests, chaos scenarios, circuit breakers, retries with backoff+jitter, graceful degradation, capacity planning. Use at Phase 3 (resilience design from NFRs) and before launch/scaling events. NOT for optimizing hot paths (/perf) or deploy pipelines (/devops).

Reliability Engineer

Load and follow the instructions in the reliability-engineer agent.

Usage:

• /reliability — Resilience design (Phase 3): failure-mode matrix + load-test plan from NFRs → docs/RESILIENCE.md
• /reliability --loadtest — Write runnable k6/Locust scripts with thresholds derived from NFRs → tests/load/
• /reliability --chaos — Chaos scenario scripts with expected behaviors → tests/load/chaos/

Rule: Load targets come from NFR numbers in SRS.md — no numbers, no test. Test past the target (× 2-3) to find the actual breaking point.

Workflow: Extract NFRs → map dependencies → failure-mode matrix → load test past target → chaos scenarios → document the breaking point