Overview
Routing should not force every request down the same path. It should balance availability, cache reuse and response time for the workload. WAFGO connects those decisions to observed behavior.
Different resources need different priorities
Small assets are sensitive to connection setup and first-byte time; large files need sustained transfer; dynamic requests depend on origin processing. Classifying traffic keeps optimization focused on the actual bottleneck.
Pages and small assets
Review request counts, connection reuse and cache keys to reduce unnecessary waiting.
Media and downloads
Measure sustained throughput, segment reuse and recovery instead of only how quickly a request starts.
Dynamic application requests
Identify requests that must reach the origin rather than hiding application latency behind unsafe shared caching.
Make health checks reflect usable service
Reachability is not the same as usable service. Choose meaningful check paths, expected responses and content conditions while keeping probe load reasonable.
- Use stable failure and recovery criteria to avoid rapid switching during transient issues.
- Verify standby data, certificates and host-header requirements before treating an origin as ready.
- Record switch times and reasons, then review effects on sessions, caches and long-lived connections.
Keep policies and recovery boundaries explicit
Automated routing still needs understandable configuration. Confirm available policies, priorities and failure conditions, with special care for stateful or write operations.
Policy groups
Group resources and entry points so a policy change does not unexpectedly affect both downloads and user transactions.
Measurement feedback
Assess errors, latency, hits and origin traffic together rather than treating one improved metric as overall success.
Controlled recovery
Keep previous configurations and assigned owners so service can recover through a defined procedure.
Implementation & checks
- 01
Classify requests
Separate static, dynamic, media and stateful traffic.
- 02
Set validation targets
Define health criteria, switch conditions and acceptable origin behavior.
- 03
Rehearse and observe
Simulate failures, recovery and route changes in a test environment.