Aller au contenu

FELoadbalancerUnhealthy#

Explanations#

Critical#

This alert rule monitors the health of servers behind a load balancer in a FlexibleEngine environment. The percentage of healthy servers is 0% for 5 minutes then this alert is triggered. This indicates that no traffic can reached your K8S cluster.

Warning#

This alert rule monitors the health of servers behind a load balancer in a FlexibleEngine environment. If the percentage of healthy servers is less than 100% for 5 minutes then this alert is triggered.

Troubleshooting hints#

To resolve this issue, complete the following tasks:

  • Connect on FE Console
  • Select the region
  • Go to Elastic Load Balance service
  • Select the concerned ELB
  • Check listeners (https; http,1080-tcp,...) to identify the unhealty backends (See https backends example below). Example of listener (https) backends
  • Check that IPs and ports are good:
    • The private IP from LB must match the K8S node IP carrying the ingress-controller pod (See nodes IPs example below). Example of nodes IP
    • The Backend port from LB must match the node port from the ingress-* service. (See backends ports example below) Example of containers (backends) ports
kubectl -n <INGRESS_NAMESPACE> get pods -o wide 
kubectl -n <INGRESS_NAMESPACE> get svc

FELoadbalancerHealthyWarning#

Explanations#

This alert rule monitors the health of servers behind a load balancer in a FlexibleEngine environment. If the percentage of healthy servers is less than 100% for 5 minutes then this alert is triggered.

Troubleshooting hints#

see above, it is the same troubleshooting hints that FELoadbalancerHealthyCritical