Added default tolerations. (#41)

aws-observability · May 22, 2024 · 32e8402 · 32e8402
1 parent cc910d3
commit 32e8402
Show file tree

Hide file tree

Showing 10 changed files with 111 additions and 2 deletions.
diff --git a/charts/amazon-cloudwatch-observability/crds/cloudwatch.aws.amazon.com_dcgmexporters.yaml b/charts/amazon-cloudwatch-observability/crds/cloudwatch.aws.amazon.com_dcgmexporters.yaml
@@ -1293,6 +1293,48 @@ spec:
                 description: TlsConfig is the raw YAML to be used as the exporter
                   TLS configuration.
                 type: string
+              tolerations:
+                description: Toleration to schedule DCGM Exporter pods.
+                  This is only relevant to daemonset, statefulset, and deployment
+                  mode
+                items:
+                  description: The pod this Toleration is attached to tolerates any
+                    taint that matches the triple <key,value,effect> using the matching
+                    operator <operator>.
+                  properties:
+                    effect:
+                      description: Effect indicates the taint effect to match. Empty
+                        means match all taint effects. When specified, allowed values
+                        are NoSchedule, PreferNoSchedule and NoExecute.
+                      type: string
+                    key:
+                      description: Key is the taint key that the toleration applies
+                        to. Empty means match all taint keys. If the key is empty,
+                        operator must be Exists; this combination means to match all
+                        values and all keys.
+                      type: string
+                    operator:
+                      description: Operator represents a key's relationship to the
+                        value. Valid operators are Exists and Equal. Defaults to Equal.
+                        Exists is equivalent to wildcard for value, so that a pod
+                        can tolerate all taints of a particular category.
+                      type: string
+                    tolerationSeconds:
+                      description: TolerationSeconds represents the period of time
+                        the toleration (which must be of effect NoExecute, otherwise
+                        this field is ignored) tolerates the taint. By default, it
+                        is not set, which means tolerate the taint forever (do not
+                        evict). Zero and negative values will be treated as 0 (evict
+                        immediately) by the system.
+                      format: int64
+                      type: integer
+                    value:
+                      description: Value is the taint value the toleration matches
+                        to. If the operator is Exists, the value should be empty,
+                        otherwise just a regular string.
+                      type: string
+                  type: object
+                type: array
               volumeMounts:
                 description: VolumeMounts represents the mount points to use in the
                   underlying collector deployment(s)

diff --git a/charts/amazon-cloudwatch-observability/crds/cloudwatch.aws.amazon.com_neuronmonitors.yaml b/charts/amazon-cloudwatch-observability/crds/cloudwatch.aws.amazon.com_neuronmonitors.yaml
@@ -1174,6 +1174,48 @@ spec:
               monitorConfig:
                 description: MonitorConfig is the raw Json to be used as monitor configuration.
                 type: string
+              tolerations:
+                description: Toleration to schedule Neuron Monitor Exporter pods.
+                  This is only relevant to daemonset, statefulset, and deployment
+                  mode
+                items:
+                  description: The pod this Toleration is attached to tolerates any
+                    taint that matches the triple <key,value,effect> using the matching
+                    operator <operator>.
+                  properties:
+                    effect:
+                      description: Effect indicates the taint effect to match. Empty
+                        means match all taint effects. When specified, allowed values
+                        are NoSchedule, PreferNoSchedule and NoExecute.
+                      type: string
+                    key:
+                      description: Key is the taint key that the toleration applies
+                        to. Empty means match all taint keys. If the key is empty,
+                        operator must be Exists; this combination means to match all
+                        values and all keys.
+                      type: string
+                    operator:
+                      description: Operator represents a key's relationship to the
+                        value. Valid operators are Exists and Equal. Defaults to Equal.
+                        Exists is equivalent to wildcard for value, so that a pod
+                        can tolerate all taints of a particular category.
+                      type: string
+                    tolerationSeconds:
+                      description: TolerationSeconds represents the period of time
+                        the toleration (which must be of effect NoExecute, otherwise
+                        this field is ignored) tolerates the taint. By default, it
+                        is not set, which means tolerate the taint forever (do not
+                        evict). Zero and negative values will be treated as 0 (evict
+                        immediately) by the system.
+                      format: int64
+                      type: integer
+                    value:
+                      description: Value is the taint value the toleration matches
+                        to. If the operator is Exists, the value should be empty,
+                        otherwise just a regular string.
+                      type: string
+                  type: object
+                type: array
               nodeSelector:
                 additionalProperties:
                   type: string

diff --git a/charts/amazon-cloudwatch-observability/templates/linux/cloudwatch-agent-daemonset.yaml b/charts/amazon-cloudwatch-observability/templates/linux/cloudwatch-agent-daemonset.yaml
@@ -111,4 +111,7 @@ spec:
     valueFrom:
       fieldRef:
         fieldPath: metadata.namespace
+  {{- with .Values.tolerations }}
+  tolerations: {{- toYaml . | nindent 2}}
+  {{- end }}
 {{- end }}
diff --git a/charts/amazon-cloudwatch-observability/templates/linux/dcgm-exporter-daemonset.yaml b/charts/amazon-cloudwatch-observability/templates/linux/dcgm-exporter-daemonset.yaml
@@ -70,4 +70,7 @@ spec:
   tlsConfig: |
     tls_server_config:
       cert_file: /etc/amazon-cloudwatch-observability-dcgm-cert/server.crt
-      key_file: /etc/amazon-cloudwatch-observability-dcgm-cert/server.key
+      key_file: /etc/amazon-cloudwatch-observability-dcgm-cert/server.key
+  {{- with .Values.tolerations }}
+  tolerations: {{- toYaml . | nindent 2}}
+  {{- end }}
diff --git a/charts/amazon-cloudwatch-observability/templates/linux/fluent-bit-daemonset.yaml b/charts/amazon-cloudwatch-observability/templates/linux/fluent-bit-daemonset.yaml
@@ -97,4 +97,7 @@ spec:
       serviceAccountName: {{ template "cloudwatch-agent.serviceAccountName" . }}
       nodeSelector:
         kubernetes.io/os: linux
+      {{- with .Values.tolerations }}
+      tolerations: {{- toYaml . | nindent 6}}
+      {{- end }}
 {{- end }}
diff --git a/charts/amazon-cloudwatch-observability/templates/linux/neuron-monitor-daemonset.yaml b/charts/amazon-cloudwatch-observability/templates/linux/neuron-monitor-daemonset.yaml
@@ -90,4 +90,7 @@ spec:
           "type": "neuron_hw_counters"
         }
       ]
-    }
+    }
+  {{- with .Values.tolerations }}
+  tolerations: {{- toYaml . | nindent 2}}
+  {{- end }}
diff --git a/charts/amazon-cloudwatch-observability/templates/operator-deployment.yaml b/charts/amazon-cloudwatch-observability/templates/operator-deployment.yaml
@@ -51,3 +51,6 @@ spec:
           secretName: {{ template "amazon-cloudwatch-observability.certificateSecretName" . }}
       nodeSelector:
         kubernetes.io/os: linux
+      {{- with .Values.tolerations }}
+      tolerations: {{- toYaml . | nindent 6}}
+      {{- end }}
diff --git a/...amazon-cloudwatch-observability/templates/windows/cloudwatch-agent-windows-daemonset.yaml b/...amazon-cloudwatch-observability/templates/windows/cloudwatch-agent-windows-daemonset.yaml
@@ -47,4 +47,7 @@ spec:
     value: "True"
   - name: RUN_AS_HOST_PROCESS_CONTAINER
     value: "True"
+  {{- with .Values.tolerations }}
+  tolerations: {{- toYaml . | nindent 2}}
+  {{- end }}
 {{- end }}
diff --git a/charts/amazon-cloudwatch-observability/templates/windows/fluent-bit-windows-daemonset.yaml b/charts/amazon-cloudwatch-observability/templates/windows/fluent-bit-windows-daemonset.yaml
@@ -70,4 +70,7 @@ spec:
       terminationGracePeriodSeconds: 10
       dnsPolicy: ClusterFirstWithHostNet
       serviceAccountName: {{ template "cloudwatch-agent.serviceAccountName" . }}
+      {{- with .Values.tolerations }}
+      tolerations: {{- toYaml . | nindent 6}}
+      {{- end }}
 {{- end }}
diff --git a/charts/amazon-cloudwatch-observability/values.yaml b/charts/amazon-cloudwatch-observability/values.yaml
@@ -24,6 +24,10 @@ gpuInstances: [ p2.xlarge, p2.8xlarge, p2.16xlarge, p3.2xlarge, p3.8xlarge, p3.1
 ## Tranium/Infrentia instance types
 neuronInstances: [ trn1.2xlarge, trn1.32xlarge, trn1n.32xlarge, inf1.xlarge, inf1.2xlarge, inf1.6xlarge, inf1.24xlarge, inf2.xlarge, inf2.8xlarge, inf2.24xlarge, inf2.48xlarge, ml.trn1.2xlarge, ml.trn1.32xlarge, ml.trn1n.32xlarge, ml.inf1.xlarge, ml.inf1.2xlarge, ml.inf1.6xlarge, ml.inf1.24xlarge, ml.inf2.xlarge, ml.inf2.8xlarge, ml.inf2.24xlarge, ml.inf2.48xlarge]
 
+## Provide default tolerations
+tolerations:
+- operator: Exists
+
 containerLogs:
   enabled: true
   fluentBit: