Update default task resources (#2377)

* Update default task resources - Defaults for task resources are too low - Upgrading the defaults to slightly higher values. Having these low defaults on cloud k8s is just wrong! fixes: #2376 Signed-off-by: Ketan Umare <[email protected]> * update single-binary Signed-off-by: Ketan Umare <[email protected]> * updated Signed-off-by: Ketan Umare <[email protected]> * Updated config Signed-off-by: Kevin Su <[email protected]> * Updated config Signed-off-by: Kevin Su <[email protected]> Co-authored-by: Kevin Su <[email protected]>
flyteorg · Apr 19, 2022 · 42cb071 · 42cb071
1 parent b351965
commit 42cb071
Show file tree

Hide file tree

Showing 11 changed files with 40 additions and 30 deletions.
diff --git a/charts/flyte-core/README.md b/charts/flyte-core/README.md
@@ -98,8 +98,8 @@ helm install gateway bitnami/contour -n flyte
 | configmap.schedulerConfig.scheduler.profilerPort | int | `10254` |  |
 | configmap.task_logs | object | `{"plugins":{"logs":{"cloudwatch-enabled":false,"kubernetes-enabled":false}}}` | Section that configures how the Task logs are displayed on the UI. This has to be changed based on your actual logging provider. Refer to [structure](https://pkg.go.dev/github.com/lyft/flyteplugins/go/tasks/logs#LogConfig) to understand how to configure various logging engines |
 | configmap.task_logs.plugins.logs.cloudwatch-enabled | bool | `false` | One option is to enable cloudwatch logging for EKS, update the region and log group accordingly |
-| configmap.task_resource_defaults | object | `{"task_resources":{"defaults":{"cpu":"100m","memory":"100Mi","storage":"5Mi"},"limits":{"cpu":2,"gpu":1,"memory":"1Gi","storage":"20Mi"}}}` | Task default resources configuration Refer to the full [structure](https://pkg.go.dev/github.com/lyft/[email protected]/pkg/runtime/interfaces#TaskResourceConfiguration). |
-| configmap.task_resource_defaults.task_resources | object | `{"defaults":{"cpu":"100m","memory":"100Mi","storage":"5Mi"},"limits":{"cpu":2,"gpu":1,"memory":"1Gi","storage":"20Mi"}}` | Task default resources parameters |
+| configmap.task_resource_defaults | object | `{"task_resources":{"defaults":{"cpu":"100m","memory":"500Mi","storage":"500Mi"},"limits":{"cpu":2,"gpu":1,"memory":"1Gi","storage":"20Mi"}}}` | Task default resources configuration Refer to the full [structure](https://pkg.go.dev/github.com/lyft/[email protected]/pkg/runtime/interfaces#TaskResourceConfiguration). |
+| configmap.task_resource_defaults.task_resources | object | `{"defaults":{"cpu":"100m","memory":"500Mi","storage":"500Mi"},"limits":{"cpu":2,"gpu":1,"memory":"1Gi","storage":"20Mi"}}` | Task default resources parameters |
 | datacatalog.affinity | object | `{}` | affinity for Datacatalog deployment |
 | datacatalog.configPath | string | `"/etc/datacatalog/config/*.yaml"` | Default regex string for searching configuration files |
 | datacatalog.enabled | bool | `true` |  |
@@ -181,7 +181,7 @@ helm install gateway bitnami/contour -n flyte
 | flytepropeller.podAnnotations | object | `{}` | Annotations for Flytepropeller pods |
 | flytepropeller.priorityClassName | string | `""` | Sets priorityClassName for propeller pod(s). |
 | flytepropeller.replicaCount | int | `1` | Replicas count for Flytepropeller deployment |
-| flytepropeller.resources | object | `{"limits":{"cpu":"200m","ephemeral-storage":"100Mi","memory":"200Mi"},"requests":{"cpu":"10m","ephemeral-storage":"50Mi","memory":"50Mi"}}` | Default resources requests and limits for Flytepropeller deployment |
+| flytepropeller.resources | object | `{"limits":{"cpu":"200m","ephemeral-storage":"100Mi","memory":"200Mi"},"requests":{"cpu":"10m","ephemeral-storage":"50Mi","memory":"100Mi"}}` | Default resources requests and limits for Flytepropeller deployment |
 | flytepropeller.serviceAccount | object | `{"annotations":{},"create":true,"imagePullSecrets":{}}` | Configuration for service accounts for FlytePropeller |
 | flytepropeller.serviceAccount.annotations | object | `{}` | Annotations for ServiceAccount attached to FlytePropeller pods |
 | flytepropeller.serviceAccount.create | bool | `true` | Should a service account be created for FlytePropeller |

diff --git a/charts/flyte-core/values-eks.yaml b/charts/flyte-core/values-eks.yaml
@@ -68,10 +68,10 @@ datacatalog:
       eks.amazonaws.com/role-arn: arn:aws:iam::{{ .Values.userSettings.accountNumber }}:role/iam-role-flyte
   resources:
     limits:
-      cpu: 500m
+      cpu: 1
       ephemeral-storage: 200Mi
     requests:
-      cpu: 50m
+      cpu: 500m
       ephemeral-storage: 200Mi
       memory: 200Mi
   service:

diff --git a/charts/flyte-core/values-gcp.yaml b/charts/flyte-core/values-gcp.yaml
@@ -26,9 +26,9 @@ flyteadmin:
     limits:
       ephemeral-storage: 2Gi
     requests:
-      cpu: 50m
+      cpu: 500m
       ephemeral-storage: 2Gi
-      memory: 200Mi
+      memory: 1G
   service:
     annotations:
       # Required for the ingress to properly route grpc traffic to grpc port
@@ -234,8 +234,8 @@ configmap:
   task_resource_defaults:
     task_resources:
       defaults:
-        cpu: 100m
-        memory: 200Mi
+        cpu: 500m
+        memory: 500Mi
         storage: 500Mi
       limits:
         storage: 2000Mi

diff --git a/charts/flyte-core/values.yaml b/charts/flyte-core/values.yaml
@@ -203,7 +203,7 @@ flytepropeller:
     requests:
       cpu: 10m
       ephemeral-storage: 50Mi
-      memory: 50Mi
+      memory: 100Mi
   cacheSizeMbs: 0
   # -- Default regex string for searching configuration files
   configPath: /etc/flyte/config/*.yaml
@@ -522,8 +522,8 @@ configmap:
     task_resources:
       defaults:
         cpu: 100m
-        memory: 100Mi
-        storage: 5Mi
+        memory: 500Mi
+        storage: 500Mi
       limits:
         cpu: 2
         memory: 1Gi

diff --git a/deployment/eks/flyte_aws_scheduler_helm_generated.yaml b/deployment/eks/flyte_aws_scheduler_helm_generated.yaml
@@ -1093,11 +1093,11 @@ spec:
         - containerPort: 10254
         resources: 
           limits:
-            cpu: 500m
+            cpu: 1
             ephemeral-storage: 200Mi
             memory: 500Mi
           requests:
-            cpu: 50m
+            cpu: 500m
             ephemeral-storage: 200Mi
             memory: 200Mi
         volumeMounts:

diff --git a/deployment/eks/flyte_helm_controlplane_generated.yaml b/deployment/eks/flyte_helm_controlplane_generated.yaml
@@ -800,11 +800,11 @@ spec:
         - containerPort: 10254
         resources: 
           limits:
-            cpu: 500m
+            cpu: 1
             ephemeral-storage: 200Mi
             memory: 500Mi
           requests:
-            cpu: 50m
+            cpu: 500m
             ephemeral-storage: 200Mi
             memory: 200Mi
         volumeMounts:

diff --git a/deployment/eks/flyte_helm_generated.yaml b/deployment/eks/flyte_helm_generated.yaml
@@ -1124,11 +1124,11 @@ spec:
         - containerPort: 10254
         resources: 
           limits:
-            cpu: 500m
+            cpu: 1
             ephemeral-storage: 200Mi
             memory: 500Mi
           requests:
-            cpu: 50m
+            cpu: 500m
             ephemeral-storage: 200Mi
             memory: 200Mi
         volumeMounts:

diff --git a/deployment/gcp/flyte_helm_controlplane_generated.yaml b/deployment/gcp/flyte_helm_controlplane_generated.yaml
@@ -162,8 +162,8 @@ data:
   task_resource_defaults.yaml: | 
     task_resources:
       defaults:
-        cpu: 100m
-        memory: 200Mi
+        cpu: 500m
+        memory: 500Mi
         storage: 500Mi
       limits:
         cpu: 2
@@ -552,7 +552,7 @@ spec:
   template:
     metadata:
       annotations:
-        configChecksum: "8ce155adcafa4029b8d11e94f999f3b51fd2eca00a4647347e700c1ba74bcd0"
+        configChecksum: "5a0ed2ad63abfcfe3ab6c31516ab0e0c3d2b54e73d2105aa75d58510e437d75"
       labels: 
         app.kubernetes.io/name: flyteadmin
         app.kubernetes.io/instance: flyte
@@ -648,9 +648,9 @@ spec:
             ephemeral-storage: 2Gi
             memory: 500Mi
           requests:
-            cpu: 50m
+            cpu: 500m
             ephemeral-storage: 2Gi
-            memory: 200Mi
+            memory: 1G
         volumeMounts:
         - mountPath: /etc/db
           name: db-pass
@@ -857,7 +857,7 @@ spec:
   template:
     metadata:
       annotations:
-        configChecksum: "8ce155adcafa4029b8d11e94f999f3b51fd2eca00a4647347e700c1ba74bcd0"
+        configChecksum: "5a0ed2ad63abfcfe3ab6c31516ab0e0c3d2b54e73d2105aa75d58510e437d75"
       labels: 
         app.kubernetes.io/name: flytescheduler
         app.kubernetes.io/instance: flyte

diff --git a/deployment/gcp/flyte_helm_generated.yaml b/deployment/gcp/flyte_helm_generated.yaml
@@ -193,8 +193,8 @@ data:
   task_resource_defaults.yaml: | 
     task_resources:
       defaults:
-        cpu: 100m
-        memory: 200Mi
+        cpu: 500m
+        memory: 500Mi
         storage: 500Mi
       limits:
         cpu: 2
@@ -884,7 +884,7 @@ spec:
   template:
     metadata:
       annotations:
-        configChecksum: "8ce155adcafa4029b8d11e94f999f3b51fd2eca00a4647347e700c1ba74bcd0"
+        configChecksum: "5a0ed2ad63abfcfe3ab6c31516ab0e0c3d2b54e73d2105aa75d58510e437d75"
       labels: 
         app.kubernetes.io/name: flyteadmin
         app.kubernetes.io/instance: flyte
@@ -980,9 +980,9 @@ spec:
             ephemeral-storage: 2Gi
             memory: 500Mi
           requests:
-            cpu: 50m
+            cpu: 500m
             ephemeral-storage: 2Gi
-            memory: 200Mi
+            memory: 1G
         volumeMounts:
         - mountPath: /etc/db
           name: db-pass
@@ -1189,7 +1189,7 @@ spec:
   template:
     metadata:
       annotations:
-        configChecksum: "8ce155adcafa4029b8d11e94f999f3b51fd2eca00a4647347e700c1ba74bcd0"
+        configChecksum: "5a0ed2ad63abfcfe3ab6c31516ab0e0c3d2b54e73d2105aa75d58510e437d75"
       labels: 
         app.kubernetes.io/name: flytescheduler
         app.kubernetes.io/instance: flyte

diff --git a/flyte.yaml b/flyte.yaml
@@ -90,3 +90,8 @@ cluster_resources:
     - projectQuotaMemory:
         value: "3000Mi"
   refresh: 5m
+task_resources:
+  defaults:
+    cpu: 500m
+    memory: 500Mi
+    storage: 500Mi
diff --git a/flyte_local.yaml b/flyte_local.yaml
@@ -90,3 +90,8 @@ flyte:
   admin:
     disableClusterResourceManager: true
     disableScheduler: true
+task_resources:
+  defaults:
+    cpu: 500m
+    memory: 500Mi
+    storage: 500Mi