A typical web request travels through 8-10 infrastructure components before reaching your application. Each hop adds latency, but also provides opportunities for optimization. The total journey can range from 50ms for cached content to 500ms+ for dynamic, uncached responses on slow networks.
The biggest wins come from eliminating entire steps: cache at the edge to skip the origin, reuse connections to skip handshakes, and compress responses to reduce transfer time. Every millisecond saved directly improves user experience and conversion rates.
⚡ Quick Decision
Optimize Connection When:
- • DNS lookups >20ms frequently
- • TLS handshakes on every request
- • Connection setup dominates latency
Optimize Infrastructure When:
- • Load balancer adds >5ms
- • API Gateway overhead >15ms
- • Multiple proxy hops exist
Optimize Payload When:
- • Response sizes >1MB
- • Mobile users predominant
- • Transfer time >100ms
💡 For implementation guides and code examples: See our technology deep dives: NGINX, Observability, HTTP/2 & HTTP/3
Request Path Latency Breakdown
Detailed latency breakdown with optimization insights for each component
Optimization Opportunities
Proven techniques to reduce request latency.
💡 Pro Tip
Combining all optimizations can reduce total latency from 290ms to under 50ms for cached content. The biggest wins come from edge caching and response compression.