Support simple ping mesh in Agent (#6120)

We introduce a new feature to measure inter-Node latency in a K8s cluster running Antrea. The feature is currently Alpha and uses the NodeLatencyMonitor FeatureGate. In addition to the FeatureGate, enablement of the feature is controlled by a new CRD, called NodeLatencyMonitor. This CRD supports at most one CR instance, which must be named "default". When the CR exists, Antrea Agents will start "pinging" each other to take latency measurements. Each Agent only stores the latest measured value (at least at the moment), we do not store time series data. We support both IPv4 and IPv6. When an oberlay is used by Antrea, the ping is sent over the tunnel (by using the gateway IP as the destination). This change does not add any functionality besides collecting latency data at each Agent. A follow-up change will take care of reporting the latency data to the Antrea Controller, so it can be consumed via an APIService. For #5514 Signed-off-by: IRONICBo <[email protected]> Signed-off-by: Asklv <[email protected]>
antrea-io · May 31, 2024 · c3103a9 · c3103a9
1 parent f7cce73
commit c3103a9
Show file tree

Hide file tree

Showing 28 changed files with 1,860 additions and 10 deletions.
diff --git a/build/charts/antrea/conf/antrea-agent.conf b/build/charts/antrea/conf/antrea-agent.conf
@@ -88,6 +88,9 @@ featureGates:
 # Enable L7FlowExporter on Pods and Namespaces to export the application layer flows such as HTTP flows.
 {{- include "featureGate" (dict "featureGates" .Values.featureGates "name" "L7FlowExporter" "default" false) }}
 
+# Enable NodeLatencyMonitor to monitor the latency between Nodes.
+{{- include "featureGate" (dict "featureGates" .Values.featureGates "name" "NodeLatencyMonitor" "default" false) }}
+
 # Name of the OpenVSwitch bridge antrea-agent will create and use.
 # Make sure it doesn't conflict with your existing OpenVSwitch bridges.
 ovsBridge: {{ .Values.ovs.bridgeName | quote }}

diff --git a/build/charts/antrea/crds/nodelatencymonitor.yaml b/build/charts/antrea/crds/nodelatencymonitor.yaml
@@ -0,0 +1,48 @@
+apiVersion: apiextensions.k8s.io/v1
+kind: CustomResourceDefinition
+metadata:
+  name: nodelatencymonitors.crd.antrea.io
+spec:
+  group: crd.antrea.io
+  versions:
+    - name: v1alpha1
+      served: true
+      storage: true
+      schema:
+        openAPIV3Schema:
+          type: object
+          required:
+            - spec
+          properties:
+            spec:
+              type: object
+              required:
+                - pingIntervalSeconds
+              properties:
+                pingIntervalSeconds:
+                  type: integer
+                  format: int32
+                  minimum: 1
+                  description: "Ping interval in seconds, must be at least 1."
+                  default: 60
+            metadata:
+              type: object
+              properties:
+                name:
+                  type: string
+                  pattern: '^default$'
+      additionalPrinterColumns:
+        - description: Specifies the interval between pings.
+          jsonPath: .spec.pingIntervalSeconds
+          name: PingIntervalSeconds
+          type: string
+        - jsonPath: .metadata.creationTimestamp
+          name: Age
+          type: date
+  scope: Cluster
+  names:
+    plural: nodelatencymonitors
+    singular: nodelatencymonitor
+    kind: NodeLatencyMonitor
+    shortNames:
+      - nlm
diff --git a/build/charts/antrea/templates/agent/clusterrole.yaml b/build/charts/antrea/templates/agent/clusterrole.yaml
@@ -174,6 +174,7 @@ rules:
       - externalippools
       - ippools
       - trafficcontrols
+      - nodelatencymonitors
     verbs:
       - get
       - watch

diff --git a/build/yamls/antrea-aks.yml b/build/yamls/antrea-aks.yml
@@ -2686,6 +2686,57 @@ spec:
       # Deprecated shortName and shall be removed in Antrea v1.14.0
       - anp
 
+---
+# Source: crds/nodelatencymonitor.yaml
+apiVersion: apiextensions.k8s.io/v1
+kind: CustomResourceDefinition
+metadata:
+  name: nodelatencymonitors.crd.antrea.io
+spec:
+  group: crd.antrea.io
+  versions:
+    - name: v1alpha1
+      served: true
+      storage: true
+      schema:
+        openAPIV3Schema:
+          type: object
+          required:
+            - spec
+          properties:
+            spec:
+              type: object
+              required:
+                - pingIntervalSeconds
+              properties:
+                pingIntervalSeconds:
+                  type: integer
+                  format: int32
+                  minimum: 1
+                  description: "Ping interval in seconds, must be at least 1."
+                  default: 60
+            metadata:
+              type: object
+              properties:
+                name:
+                  type: string
+                  pattern: '^default$'
+      additionalPrinterColumns:
+        - description: Specifies the interval between pings.
+          jsonPath: .spec.pingIntervalSeconds
+          name: PingIntervalSeconds
+          type: string
+        - jsonPath: .metadata.creationTimestamp
+          name: Age
+          type: date
+  scope: Cluster
+  names:
+    plural: nodelatencymonitors
+    singular: nodelatencymonitor
+    kind: NodeLatencyMonitor
+    shortNames:
+      - nlm
+
 ---
 # Source: crds/supportbundlecollection.yaml
 apiVersion: apiextensions.k8s.io/v1
@@ -3624,6 +3675,9 @@ data:
     # Enable L7FlowExporter on Pods and Namespaces to export the application layer flows such as HTTP flows.
     #  L7FlowExporter: false
 
+    # Enable NodeLatencyMonitor to monitor the latency between Nodes.
+    #  NodeLatencyMonitor: false
+
     # Name of the OpenVSwitch bridge antrea-agent will create and use.
     # Make sure it doesn't conflict with your existing OpenVSwitch bridges.
     ovsBridge: "br-int"
@@ -4259,6 +4313,7 @@ rules:
       - externalippools
       - ippools
       - trafficcontrols
+      - nodelatencymonitors
     verbs:
       - get
       - watch
@@ -4920,7 +4975,7 @@ spec:
         kubectl.kubernetes.io/default-container: antrea-agent
         # Automatically restart Pods with a RollingUpdate if the ConfigMap changes
         # See https://helm.sh/docs/howto/charts_tips_and_tricks/#automatically-roll-deployments
-        checksum/config: 30843b57762c91dfcffb560917191e3bc7e662c06552759bac2a173bc060b82c
+        checksum/config: 47a8888bb99a5b1a08dea61e9315bacf613d869d718712ad0eb9964bb73dc0ec
       labels:
         app: antrea
         component: antrea-agent
@@ -5158,7 +5213,7 @@ spec:
       annotations:
         # Automatically restart Pod if the ConfigMap changes
         # See https://helm.sh/docs/howto/charts_tips_and_tricks/#automatically-roll-deployments
-        checksum/config: 30843b57762c91dfcffb560917191e3bc7e662c06552759bac2a173bc060b82c
+        checksum/config: 47a8888bb99a5b1a08dea61e9315bacf613d869d718712ad0eb9964bb73dc0ec
       labels:
         app: antrea
         component: antrea-controller

diff --git a/build/yamls/antrea-crds.yml b/build/yamls/antrea-crds.yml
@@ -2667,6 +2667,55 @@ spec:
 ---
 apiVersion: apiextensions.k8s.io/v1
 kind: CustomResourceDefinition
+metadata:
+  name: nodelatencymonitors.crd.antrea.io
+spec:
+  group: crd.antrea.io
+  versions:
+    - name: v1alpha1
+      served: true
+      storage: true
+      schema:
+        openAPIV3Schema:
+          type: object
+          required:
+            - spec
+          properties:
+            spec:
+              type: object
+              required:
+                - pingIntervalSeconds
+              properties:
+                pingIntervalSeconds:
+                  type: integer
+                  format: int32
+                  minimum: 1
+                  description: "Ping interval in seconds, must be at least 1."
+                  default: 60
+            metadata:
+              type: object
+              properties:
+                name:
+                  type: string
+                  pattern: '^default$'
+      additionalPrinterColumns:
+        - description: Specifies the interval between pings.
+          jsonPath: .spec.pingIntervalSeconds
+          name: PingIntervalSeconds
+          type: string
+        - jsonPath: .metadata.creationTimestamp
+          name: Age
+          type: date
+  scope: Cluster
+  names:
+    plural: nodelatencymonitors
+    singular: nodelatencymonitor
+    kind: NodeLatencyMonitor
+    shortNames:
+      - nlm
+---
+apiVersion: apiextensions.k8s.io/v1
+kind: CustomResourceDefinition
 metadata:
   name: supportbundlecollections.crd.antrea.io
 spec:

diff --git a/build/yamls/antrea-eks.yml b/build/yamls/antrea-eks.yml
@@ -2686,6 +2686,57 @@ spec:
       # Deprecated shortName and shall be removed in Antrea v1.14.0
       - anp
 
+---
+# Source: crds/nodelatencymonitor.yaml
+apiVersion: apiextensions.k8s.io/v1
+kind: CustomResourceDefinition
+metadata:
+  name: nodelatencymonitors.crd.antrea.io
+spec:
+  group: crd.antrea.io
+  versions:
+    - name: v1alpha1
+      served: true
+      storage: true
+      schema:
+        openAPIV3Schema:
+          type: object
+          required:
+            - spec
+          properties:
+            spec:
+              type: object
+              required:
+                - pingIntervalSeconds
+              properties:
+                pingIntervalSeconds:
+                  type: integer
+                  format: int32
+                  minimum: 1
+                  description: "Ping interval in seconds, must be at least 1."
+                  default: 60
+            metadata:
+              type: object
+              properties:
+                name:
+                  type: string
+                  pattern: '^default$'
+      additionalPrinterColumns:
+        - description: Specifies the interval between pings.
+          jsonPath: .spec.pingIntervalSeconds
+          name: PingIntervalSeconds
+          type: string
+        - jsonPath: .metadata.creationTimestamp
+          name: Age
+          type: date
+  scope: Cluster
+  names:
+    plural: nodelatencymonitors
+    singular: nodelatencymonitor
+    kind: NodeLatencyMonitor
+    shortNames:
+      - nlm
+
 ---
 # Source: crds/supportbundlecollection.yaml
 apiVersion: apiextensions.k8s.io/v1
@@ -3624,6 +3675,9 @@ data:
     # Enable L7FlowExporter on Pods and Namespaces to export the application layer flows such as HTTP flows.
     #  L7FlowExporter: false
 
+    # Enable NodeLatencyMonitor to monitor the latency between Nodes.
+    #  NodeLatencyMonitor: false
+
     # Name of the OpenVSwitch bridge antrea-agent will create and use.
     # Make sure it doesn't conflict with your existing OpenVSwitch bridges.
     ovsBridge: "br-int"
@@ -4259,6 +4313,7 @@ rules:
       - externalippools
       - ippools
       - trafficcontrols
+      - nodelatencymonitors
     verbs:
       - get
       - watch
@@ -4920,7 +4975,7 @@ spec:
         kubectl.kubernetes.io/default-container: antrea-agent
         # Automatically restart Pods with a RollingUpdate if the ConfigMap changes
         # See https://helm.sh/docs/howto/charts_tips_and_tricks/#automatically-roll-deployments
-        checksum/config: 30843b57762c91dfcffb560917191e3bc7e662c06552759bac2a173bc060b82c
+        checksum/config: 47a8888bb99a5b1a08dea61e9315bacf613d869d718712ad0eb9964bb73dc0ec
       labels:
         app: antrea
         component: antrea-agent
@@ -5159,7 +5214,7 @@ spec:
       annotations:
         # Automatically restart Pod if the ConfigMap changes
         # See https://helm.sh/docs/howto/charts_tips_and_tricks/#automatically-roll-deployments
-        checksum/config: 30843b57762c91dfcffb560917191e3bc7e662c06552759bac2a173bc060b82c
+        checksum/config: 47a8888bb99a5b1a08dea61e9315bacf613d869d718712ad0eb9964bb73dc0ec
       labels:
         app: antrea
         component: antrea-controller
-Original file line number
+Diff line change
@@ Expand Up / @@ -174,6 +174,7 @@ rules: @@
           - externalippools
           - ippools
           - trafficcontrols
+          - nodelatencymonitors
         verbs:
           - get
           - watch
@@ Expand Down @@