Prompt
trace slow requests
Latest observation
Nov 21, 2026
OpenAI APIWeb search: off
To trace slow requests, I need a bit more context—what stack are you using?
Common options:
- Web app logs: add request timing middleware and log slow requests
- APM tools: Datadog, New Relic, AppDynamics, Elastic APM
- Tracing: OpenTelemetry + Jaeger/Tempo/Zipkin
- Infra: nginx / load balancer access logs, database slow query logs
If you want a quick generic approach, here’s the usual workflow:
-
Identify the slow request
- log URL, method, status, duration, request id, user id
-
Break down the time
- app code
- database queries
- external API calls
- cache misses
- queue waits
-
Add tracing
- propagate a trace/request id through all services
- instrument important spans:
- controller
- service layer
- DB query
- outbound HTTP call
-
Check dependencies
- slow SQL
- timeouts/retries
- lock contention
- thread pool saturation
- GC / CPU / memory pressure
-
Correlate with logs
- filter by request id / trace id
- inspect spans around the slow request
If you share your language/framework (for example: Node, Python, Java, Go, .NET) and where the request is slow, I can show you exactly how to trace it.