Add metrics-server to list of platform namespaces

Testing:
* Deployed ISO with changes.
* Configured kube-cpu-mgr-policy=static.
* Verified that metrics-server were running on platform CPUs.

Partial-Bug: 1964503

Signed-off-by: Thiago Miranda <ThiagoOliveira.Miranda@windriver.com>
Change-Id: I9bebe9ec27fcd70e89a4cae52bfacde993f958eb
This commit is contained in:
Thiago Miranda 2022-03-10 12:04:30 -05:00
parent acdd869a58
commit 636e5352d9
3 changed files with 6 additions and 6 deletions

View File

@ -9,7 +9,7 @@ to isolate them from the shared pool of CPUs.
Infrastructure pods include any pods that belong to the kube-system,
armada, cert-manager, vault, platform-deployment-manager, portieris,
notification or flux-helm namespaces.
notification, flux-helm or metrics-server namespaces.
The implementation is a bit simplistic, it is assumed that the
"reserved" cpuset is large enough to handle all infrastructure pods
@ -33,7 +33,7 @@ index 94d81796132..b48a5b997a3 100644
+// Define namespaces used by platform infrastructure pods
+var infraNamespaces = [...]string{
+ "kube-system", "armada", "cert-manager", "platform-deployment-manager", "portieris", "vault", "notification", "flux-helm",
+ "kube-system", "armada", "cert-manager", "platform-deployment-manager", "portieris", "vault", "notification", "flux-helm", "metrics-server",
+}
+
// staticPolicy is a CPU manager policy that does not change CPU

View File

@ -9,7 +9,7 @@ to isolate them from the shared pool of CPUs.
Infrastructure pods include any pods that belong to the kube-system,
armada, cert-manager, vault, platform-deployment-manager, portieris,
notification or flux-helm namespaces.
notification, flux-helm or metrics-server namespaces.
The implementation is a bit simplistic, it is assumed that the
"reserved" cpuset is large enough to handle all infrastructure pods
@ -33,7 +33,7 @@ index 94d81796132..b48a5b997a3 100644
+// Define namespaces used by platform infrastructure pods
+var infraNamespaces = [...]string{
+ "kube-system", "armada", "cert-manager", "platform-deployment-manager", "portieris", "vault", "notification", "flux-helm",
+ "kube-system", "armada", "cert-manager", "platform-deployment-manager", "portieris", "vault", "notification", "flux-helm", "metrics-server",
+}
+
// staticPolicy is a CPU manager policy that does not change CPU

View File

@ -9,7 +9,7 @@ to isolate them from the shared pool of CPUs.
Infrastructure pods include any pods that belong to the kube-system,
armada, cert-manager, vault, platform-deployment-manager, portieris,
notification or flux-helm namespaces.
notification, flux-helm or metrics-server namespaces.
The implementation is a bit simplistic, it is assumed that the
"reserved" cpuset is large enough to handle all infrastructure pods
@ -33,7 +33,7 @@ index 89625fc6..492457b1 100644
+// Define namespaces used by platform infrastructure pods
+var infraNamespaces = [...]string{
+ "kube-system", "armada", "cert-manager", "platform-deployment-manager", "portieris", "vault", "notification", "flux-helm",
+ "kube-system", "armada", "cert-manager", "platform-deployment-manager", "portieris", "vault", "notification", "flux-helm", "metrics-server",
+}
+
// staticPolicy is a CPU manager policy that does not change CPU