You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
On x86 system running container-kepler v0.7.10 bare-metal:
root# curl localhost:8888/metrics | grep kepler | wc -l
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 163k 0 163k 0 0 15.9M 0 --:--:-- --:--:-- --:--:-- 15.9M 1213
Pick a process and search for metrics per PID (ie. PID 986 = NetworkManager)
root# cat HOLD.txt | grep 986 | wc -l
25
What happened?
On x86 system running container-kepler v0.7.10 bare-metal:
root# curl localhost:8888/metrics | grep kepler | wc -l
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 163k 0 163k 0 0 15.9M 0 --:--:-- --:--:-- --:--:-- 15.9M
1213
Pick a process and search for metrics per PID (ie. PID 986 = NetworkManager)
root# cat HOLD.txt | grep 986 | wc -l
25
Break down record count by record-type:
root# cat HOLD.txt | grep 986 | grep rapl | wc -l
12
root# cat HOLD.txt | grep 986 | grep cgroup | wc -l
4
root@# cat HOLD.txt | grep 986 | grep bpf | wc -l
9
Twenty-five records per PID. That's a lot of metrics and a large payload.
What did you expect to happen?
Raising concern on potential overhead of recording so many metrics per PID
How can we reproduce it (as minimally and precisely as possible)?
See cmd sequence above
Anything else we need to know?
No response
Kepler image tag
Kubernetes version
Cloud provider or bare metal
OS version
Install tools
Kepler deployment config
For on kubernetes:
For standalone:
put your Kepler command argument here
Container runtime (CRI) and version (if applicable)
Related plugins (CNI, CSI, ...) and versions (if applicable)
The text was updated successfully, but these errors were encountered: