Hi Andrew,
Load Balancers on AWS EKS does not support UDP protocol what forces me to set up a graylog process on each node of the k8s cluster, which in turn is waste of cpu and ram
My idea is to replace the graylog processes with grayproxy because of its smaller footprint.
For Kubernetes know when the process is ready to get traffic and also know if it continues healthy it requires to run a readiness and liveness probes, be it run a command (exit code=0 success), check if a tcp socket is listening (if it opens=success) or check if an http port is listening (if 200>=statuscode<=400 is success)
Have you thought on adding that kind of probes to the process? I am going to write a Helm Chart to setup Grayproxy on K8S and would be a great addition.
Hi Andrew,
Load Balancers on AWS EKS does not support UDP protocol what forces me to set up a graylog process on each node of the k8s cluster, which in turn is waste of cpu and ram
My idea is to replace the graylog processes with grayproxy because of its smaller footprint.
For Kubernetes know when the process is ready to get traffic and also know if it continues healthy it requires to run a readiness and liveness probes, be it run a command (exit code=0 success), check if a tcp socket is listening (if it opens=success) or check if an http port is listening (if 200>=statuscode<=400 is success)
Have you thought on adding that kind of probes to the process? I am going to write a Helm Chart to setup Grayproxy on K8S and would be a great addition.