Find upstream timeouts in logs
Find backend timeouts and requests the client dropped. Note: 499 means the client closed the connection before upstream replied
kubectl logs -n ingress-nginx deploy/ingress-nginx-controller --tail=2000 | grep -iE 'upstream timed out|504|499'
- tool
- ingress-nginx7 commands
More ingress-nginx commands
-
kubectl annotate ingress <name> -n <ns> nginx.ingress.kubernetes.io/canary=true nginx.ingress.kubernetes.io/canary-weight="20"Send 20% of traffic to a canary ingress. You need a main ingress on the same host and path
-
kubectl annotate ingress <name> -n <ns> nginx.ingress.kubernetes.io/limit-rps="10" nginx.ingress.kubernetes.io/proxy-body-size="50m"Limit requests per second per client IP and raise the upload limit. Burst is 5x rps by default
-
kubectl annotate ingress <name> -n <ns> nginx.ingress.kubernetes.io/rewrite-target='/$2' nginx.ingress.kubernetes.io/use-regex='true'Strip a path prefix with a regex capture. The path must use (/|$)(.*) so that $2 works
-
kubectl get ingressclass -o custom-columns='NAME:.metadata.name,DEFAULT:.metadata.annotations.ingressclass\.kubernetes\.io/is-default-class'Show controllers and which IngressClass is the default. Only one should be marked as default
-
kubectl exec deploy/ingress-nginx-controller -n ingress-nginx -- nginx -T | lessPrint the full live nginx.conf with all server blocks the controller really generated