What problem does it solve?
This Skill helps you reduce latency and improve resilience in VTEX IO backends by applying production-ready application-level caching, deduplication, and parallelization patterns instead of relying only on edge caching.
Core Features & Use Cases
- Multi-layer caching: Use bounded in-process LRU for fast hot reads, optionally backed by shared VBase with stale-while-revalidate for lower tail latency.
- Request-safe deduplication: Prevent duplicate upstream calls within a resolver chain using ctx.state flags and explicit client/context wiring (e.g., ctx.clients).
- Performance-safe execution: Parallelize independent upstream calls with phased Promise.all and enforce explicit timeouts and graceful degradation for reliability.
- Correctness guardrails: Avoid caching transactional real-time data (order forms, cart simulations, payment responses) and apply strict constraints for idempotency-critical VBase writes.
Quick Start
Use vtex-io-application-performance to audit your VTEX IO resolvers and refactor hot-path logic to add tenant-scoped LRU/VBase caching for reference data, parallelize independent client calls, and ensure timeouts and non-cached transactional data are enforced.
