github/kubernetes
1
0
Fork 0
mirror of https://github.com/k3s-io/kubernetes.git synced 2026-01-21 13:49:13 +00:00
Code Issues Projects Releases Wiki Activity

Merge pull request #126287 from devppratik/121793-update-node-monitor-grace-period

Browse Source 
node: Update Node Monitor Grace Period default duration to 50s
This commit is contained in:
Kubernetes Prow Robot
2024-08-13 21:03:16 -07:00
committed by GitHub
parent cc4ed1d3fa f8bf6b97b8
commit bb7411120a
7 changed files with 19 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
pkg/controller/nodelifecycle/config/types.go
View File

@@ -32,7 +32,8 @@ type NodeLifecycleControllerConfiguration struct {
// NodeMonitorGracePeriod is the amount of time which we allow a running node to be
// unresponsive before marking it unhealthy. Must be N times more than kubelet's
// nodeStatusUpdateFrequency, where N means number of retries allowed for kubelet
// to post node status.
// to post node status. This value should also be greater than the sum of
// HTTP2_PING_TIMEOUT_SECONDS and HTTP2_READ_IDLE_TIMEOUT_SECONDS.
NodeMonitorGracePeriod metav1.Duration
// secondaryNodeEvictionRate is implicitly overridden to 0 for clusters smaller than or equal to largeClusterSizeThreshold
LargeClusterSizeThreshold int32

7
pkg/controller/nodelifecycle/config/v1alpha1/defaults.go
View File

@@ -37,8 +37,13 @@ func RecommendedDefaultNodeLifecycleControllerConfiguration(obj *kubectrlmgrconf
if obj.PodEvictionTimeout == zero {
obj.PodEvictionTimeout = metav1.Duration{Duration: 5 * time.Minute}
}
// NodeMonitorGracePeriod is set to a default value of 50 seconds.
// This value should be greater than the sum of HTTP2_PING_TIMEOUT_SECONDS (30s)
// and HTTP2_READ_IDLE_TIMEOUT_SECONDS (15s) from the http2 health check
// to ensure that the server has adequate time to handle slow or idle connections
// properly before marking a node as unhealthy.
if obj.NodeMonitorGracePeriod == zero {
obj.NodeMonitorGracePeriod = metav1.Duration{Duration: 40 * time.Second}
obj.NodeMonitorGracePeriod = metav1.Duration{Duration: 50 * time.Second}
}
if obj.NodeStartupGracePeriod == zero {
obj.NodeStartupGracePeriod = metav1.Duration{Duration: 60 * time.Second}

6
pkg/controller/nodelifecycle/node_lifecycle_controller.go
View File

@@ -284,7 +284,11 @@ type Controller struct {
// be less than the node health signal update frequency, since there will
// only be fresh values from Kubelet at an interval of node health signal
// update frequency.
// 2. nodeMonitorGracePeriod can't be too large for user experience - larger
// 2. nodeMonitorGracePeriod should be greater than the sum of HTTP2_PING_TIMEOUT_SECONDS (30s)
// and HTTP2_READ_IDLE_TIMEOUT_SECONDS (15s) from the http2 health check
// to ensure that the server has adequate time to handle slow or idle connections
// properly before marking a node as unhealthy.
// 3. nodeMonitorGracePeriod can't be too large for user experience - larger
// value takes longer for user to see up-to-date node health.
nodeMonitorGracePeriod time.Duration

2
pkg/controller/nodelifecycle/node_lifecycle_controller_test.go
View File

@@ -52,7 +52,7 @@ import (
)
const (
testNodeMonitorGracePeriod = 40 * time.Second
testNodeMonitorGracePeriod = 50 * time.Second
testNodeStartupGracePeriod = 60 * time.Second
testNodeMonitorPeriod = 5 * time.Second
testRateLimiterQPS = float32(100000)