← ClaudeAtlas

hightriadlisted

Design and review systems for high concurrency, high performance, and high availability. Use when planning architecture, scaling strategies, SLOs/SLIs, load testing, resilience, failover, capacity planning, caching, data access optimization, or incident-readiness for production systems.
JWCodeWrote/Agent_Skills_Plugin · ★ 3 · AI & Automation · score 55
Install: claude install-skill JWCodeWrote/Agent_Skills_Plugin
# HighTriad Build professional, production-grade system designs that balance high concurrency, high performance, and high availability. ## Core Workflow 1. **Clarify requirements** Collect workload shape, critical paths, and failure tolerance. Ask for absolute targets: RPS/QPS, p95/p99 latency, peak traffic, growth rate, error budget, RTO/RPO, data consistency needs. 2. **Define SLIs/SLOs** Choose 3 to 5 primary SLIs and map them to explicit SLOs. Prefer latency percentiles, availability, throughput, and freshness over averages. 3. **Model the system** Sketch request flow and identify bottlenecks across compute, network, storage, and dependency call chains. Enumerate concurrency boundaries: queues, pools, locks, partitions, and external rate limits. 4. **Design for scale** Select scaling axis: horizontal, vertical, data partitioning, or event-driven async. Define partitioning keys, load balancing strategy, and caching boundaries. 5. **Design for performance** Minimize critical path length, reduce tail latency, and cut remote calls. Choose data access patterns, indexing, caching tiers, and compression tradeoffs. 6. **Design for availability** Add redundancy, fault isolation, and graceful degradation. Define failover paths, health checks, circuit breakers, and data durability strategy. 7. **Validate with tests** Create load, stress, soak, and chaos test plans aligned to SLOs. Plan rollback and mitigation steps for regression risk. 8. **Operationalize** Define observab