fix: give higher memory overhead percent

Signed-off-by: jwcesign <[email protected]>

charts/karpenter/values.yaml

@@ -64,7 +64,7 @@ controller:
     # -- The external kubernetes cluster id for new nodes to connect with.
     clusterID: ""
     # -- The VM memory overhead as a percent that will be subtracted from the total memory for all instance types. The value of `0.075` equals to 7.5%.
-    vmMemoryOverheadPercent: 0.065
+    vmMemoryOverheadPercent: 0.075
     # -- The maximum length of a batch window. The longer this is, the more pods we can consider for provisioning at one
     # time which usually results in fewer but larger nodes.
     batchMaxDuration: 10s

pkg/operator/options/options.go

@@ -44,7 +44,7 @@ type Options struct {
 func (o *Options) AddFlags(fs *coreoptions.FlagSet) {
 	fs.StringVar(&o.ClusterID, "cluster-id", env.WithDefaultString("CLUSTER_ID", ""), "The external kubernetes cluster id for new nodes to connect with.")
 	// TODO: for different OS, the overhead is different, find a way to fix this.
-	fs.Float64Var(&o.VMMemoryOverheadPercent, "vm-memory-overhead-percent", utils.WithDefaultFloat64("VM_MEMORY_OVERHEAD_PERCENT", 0.065), "The VM memory overhead as a percent that will be subtracted from the total memory for all instance types.")
+	fs.Float64Var(&o.VMMemoryOverheadPercent, "vm-memory-overhead-percent", utils.WithDefaultFloat64("VM_MEMORY_OVERHEAD_PERCENT", 0.075), "The VM memory overhead as a percent that will be subtracted from the total memory for all instance types.")
 	fs.BoolVar(&o.Interruption, "interruption", env.WithDefaultBool("INTERRUPTION", true), "Enable interruption handling.")
 	fs.BoolVar(&o.TelemetryShare, "telemetry-share", env.WithDefaultBool("TELEMETRY_SHARE", true), "Enable telemetry sharing.")
 	fs.IntVar(&o.APGCreationQPS, "apg-qps", int(env.WithDefaultInt64("APG_CREATION_QPS", 100)), "The QPS limit for creating AutoProvisionGroup.")