Refactor cpu limit as nullable field

caraml-dev · May 24, 2024 · 8535c51 · 8535c51
1 parent a255f43
commit 8535c51
Show file tree

Hide file tree

Showing 3 changed files with 9 additions and 8 deletions.
diff --git a/api/cluster/resource/templater.go b/api/cluster/resource/templater.go
@@ -198,7 +198,7 @@ func (t *InferenceServiceTemplater) createPredictorSpec(modelService *models.Ser
 	limits := map[corev1.ResourceName]resource.Quantity{}
 
 	// Set cpu resource limits automatically if they have not been set
-	if modelService.ResourceRequest.CPULimit.IsZero() {
+	if modelService.ResourceRequest.CPULimit == nil || modelService.ResourceRequest.CPULimit.IsZero() {
 		if t.deploymentConfig.UserContainerCPULimitRequestFactor != 0 {
 			limits[corev1.ResourceCPU] = ScaleQuantity(
 				modelService.ResourceRequest.CPURequest, t.deploymentConfig.UserContainerCPULimitRequestFactor,
@@ -214,7 +214,7 @@ func (t *InferenceServiceTemplater) createPredictorSpec(modelService *models.Ser
 			envVars = models.MergeEnvVars(ParseEnvVars(t.deploymentConfig.DefaultEnvVarsWithoutCPULimits), envVars)
 		}
 	} else {
-		limits[corev1.ResourceCPU] = modelService.ResourceRequest.CPULimit
+		limits[corev1.ResourceCPU] = *modelService.ResourceRequest.CPULimit
 	}
 
 	if t.deploymentConfig.UserContainerMemoryLimitRequestFactor != 0 {
@@ -406,7 +406,7 @@ func (t *InferenceServiceTemplater) createTransformerSpec(
 	// Set resource limits to request * userContainerCPULimitRequestFactor or UserContainerMemoryLimitRequestFactor
 	limits := map[corev1.ResourceName]resource.Quantity{}
 	// Set cpu resource limits automatically if they have not been set
-	if transformer.ResourceRequest.CPULimit.IsZero() {
+	if transformer.ResourceRequest.CPULimit == nil || transformer.ResourceRequest.CPULimit.IsZero() {
 		if t.deploymentConfig.UserContainerCPULimitRequestFactor != 0 {
 			limits[corev1.ResourceCPU] = ScaleQuantity(
 				transformer.ResourceRequest.CPURequest, t.deploymentConfig.UserContainerCPULimitRequestFactor,
@@ -422,7 +422,7 @@ func (t *InferenceServiceTemplater) createTransformerSpec(
 			envVars = models.MergeEnvVars(ParseEnvVars(t.deploymentConfig.DefaultEnvVarsWithoutCPULimits), envVars)
 		}
 	} else {
-		limits[corev1.ResourceCPU] = transformer.ResourceRequest.CPULimit
+		limits[corev1.ResourceCPU] = *transformer.ResourceRequest.CPULimit
 	}
 
 	if t.deploymentConfig.UserContainerMemoryLimitRequestFactor != 0 {

diff --git a/api/cluster/resource/templater_test.go b/api/cluster/resource/templater_test.go
@@ -93,12 +93,13 @@ var (
 		MemoryRequest: resource.MustParse("1Gi"),
 	}
 
+	cpuLimit                          = resource.MustParse("8")
 	userResourceRequestsWithCPULimits = &models.ResourceRequest{
 		MinReplica:    1,
 		MaxReplica:    10,
 		CPURequest:    resource.MustParse("1"),
 		MemoryRequest: resource.MustParse("1Gi"),
-		CPULimit:      resource.MustParse("8"),
+		CPULimit:      &cpuLimit,
 	}
 
 	expUserResourceRequests = corev1.ResourceRequirements{
@@ -118,7 +119,7 @@ var (
 			corev1.ResourceMemory: userResourceRequestsWithCPULimits.MemoryRequest,
 		},
 		Limits: corev1.ResourceList{
-			corev1.ResourceCPU:    userResourceRequestsWithCPULimits.CPULimit,
+			corev1.ResourceCPU:    *userResourceRequestsWithCPULimits.CPULimit,
 			corev1.ResourceMemory: ScaleQuantity(userResourceRequestsWithCPULimits.MemoryRequest, 2),
 		},
 	}
@@ -4627,7 +4628,7 @@ func TestCreateTransformerSpec(t *testing.T) {
 						MaxReplica:    1,
 						CPURequest:    cpuRequest,
 						MemoryRequest: memoryRequest,
-						CPULimit:      customCPULimit,
+						CPULimit:      &customCPULimit,
 					},
 					EnvVars: models.EnvVars{
 						{Name: transformerpkg.JaegerCollectorURL, Value: "NEW_HOST"}, // test user overwrite

diff --git a/api/models/resource_request.go b/api/models/resource_request.go
@@ -30,7 +30,7 @@ type ResourceRequest struct {
 	// CPU request of inference service
 	CPURequest resource.Quantity `json:"cpu_request"`
 	// CPU limit of inference service
-	CPULimit resource.Quantity `json:"cpu_limit"`
+	CPULimit *resource.Quantity `json:"cpu_limit,omitempty"`
 	// Memory request of inference service
 	MemoryRequest resource.Quantity `json:"memory_request"`
 	// GPU name