Audit the cgroups in Talos and resource reservation #7081
Assignees
Comments
|
This issue is stale because it has been open 180 days with no activity. Remove stale label or comment or this will be closed in 7 days. |
|
This issue was closed because it has been stalled for 7 days with no activity. |
|
Still valid, need a stress test |
smira
added a commit
to smira/talos
that referenced
this issue
Sep 19, 2024
Fixes: siderolabs#7081 Review all reservations and limits set, test under stress load (using both memory and CPU). The goal: system components (Talos itself) and runtime (kubelet, CRI) should survive under extreme resource starvation (workloads consuming all CPU/memory). Uses siderolabs#9337 to visualize changes, but doesn't depend on it. Signed-off-by: Andrey Smirnov <andrey.smirnov@siderolabs.com>
smira
added a commit
to smira/talos
that referenced
this issue
Sep 19, 2024
Fixes: siderolabs#7081 Review all reservations and limits set, test under stress load (using both memory and CPU). The goal: system components (Talos itself) and runtime (kubelet, CRI) should survive under extreme resource starvation (workloads consuming all CPU/memory). Uses siderolabs#9337 to visualize changes, but doesn't depend on it. Signed-off-by: Andrey Smirnov <andrey.smirnov@siderolabs.com>
smira
added a commit
to smira/talos
that referenced
this issue
Sep 20, 2024
Fixes: siderolabs#7081 Review all reservations and limits set, test under stress load (using both memory and CPU). The goal: system components (Talos itself) and runtime (kubelet, CRI) should survive under extreme resource starvation (workloads consuming all CPU/memory). Uses siderolabs#9337 to visualize changes, but doesn't depend on it. Signed-off-by: Andrey Smirnov <andrey.smirnov@siderolabs.com>
smira
added a commit
to smira/talos
that referenced
this issue
Sep 20, 2024
Fixes: siderolabs#7081 Review all reservations and limits set, test under stress load (using both memory and CPU). The goal: system components (Talos itself) and runtime (kubelet, CRI) should survive under extreme resource starvation (workloads consuming all CPU/memory). Uses siderolabs#9337 to visualize changes, but doesn't depend on it. Signed-off-by: Andrey Smirnov <andrey.smirnov@siderolabs.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Do we need to reserve some CPU for
/initcgroup (machined)?Do we want to have some default reservation for the kubelet cgroup?
Can we easily gather some resource consumption data on the dashboard?
The text was updated successfully, but these errors were encountered: