vtex-io-application-performance

Reduce VTEX IO backend latency by implementing layered caching using in-process LRU for hot keys and VBase-backed shared caching with stale-while-revalidate to minimize tail latency and upstream calls.

Stale-while-revalidate caching in VTEX IO serves stale data from VBase while asynchronously refreshing it, minimizing tail latency and improving resilience during upstream delays or outages.

Parallelize outbound calls in Node.js VTEX IO apps by executing independent I/O operations concurrently and deduplicating resolver chains, which reduces duplicate work and improves throughput.

Yes, tenancy-aware caching supports per-tenant isolation by using per-entity VBase keys and per-account workspace configurations, ensuring cached data remains strictly separated across tenants.

Safely load AppSettings in VTEX IO by fetching configuration at startup time and applying per-request context patterns, avoiding repeated config reloading and ensuring safe configuration loading.

Use in-process LRU cache for hot keys to reduce repeated upstream calls, and use VBase-backed shared cache for broader persistence across instances with stale-while-revalidate for resilience.