Closed
Description
Although we have pod level detail within k8s, there are times it may be necessary to understanding the process level utilisation of CPU and memory within the pod/container to assess inefficiencies.
At present, one would need to instruct the customer to access the pod to run various commands which can be prone to error.
Describe the solution you'd like
The objective of this issue is to explore collecting pod specific process information such as ps output, memory etc.
Activity
cpuengr949 commentedon Sep 5, 2024
Some of the commands that have been asked to run are below, it would be nice to have all of this in the support package if needed.
cat /sys/fs/cgroup/memory.stat
from within a Ingress Controller podfor p in $(pidof nginx nginx-ingress); do sudo pmap ${p} -X; done
from the k8s node running the Ingress Controller podsfor p in $(pidof nginx nginx-ingress); do cat /proc/${p}/status; done
since azure was unable to run the command aboveCollect the following information from the API endpoint from the pods:
mrajagopal commentedon Nov 19, 2024
Our plan is to deploy a custom container as a sidecar in the pod to fetch the necessary data.
mrajagopal commentedon Dec 3, 2024
f5-debugger
container tonginx-ingress
container, we can now get the API output usingcurl
andjq
: