rename criticality to priority (#1363)

Author: ahg-g

apix/v1alpha2/inferenceobjective_types.go

@@ -27,7 +27,7 @@ import (
 // +kubebuilder:storageversion
 // +kubebuilder:printcolumn:name="Model Name",type=string,JSONPath=`.spec.modelName`
 // +kubebuilder:printcolumn:name="Inference Pool",type=string,JSONPath=`.spec.poolRef.name`
-// +kubebuilder:printcolumn:name="Criticality",type=string,JSONPath=`.spec.criticality`
+// +kubebuilder:printcolumn:name="Priority",type=string,JSONPath=`.spec.priority`
 // +kubebuilder:printcolumn:name="Age",type=date,JSONPath=`.metadata.creationTimestamp`
 // +genclient
 type InferenceObjective struct {
@@ -64,20 +64,20 @@ type InferenceObjectiveList struct {
 // condition, one will be selected at random.
 type InferenceObjectiveSpec struct {
 
-	// Criticality defines how important it is to serve the request compared to other requests in the same pool.
-	// Criticality is an integer value that defines the priority of the request.
+	// Priority defines how important it is to serve the request compared to other requests in the same pool.
+	// Priority is an integer value that defines the priority of the request.
 	// The higher the value, the more critical the request is; negative values _are_ allowed.
 	// No default value is set for this field, allowing for future additions of new fields that may 'one of' with this field.
 	// However, implementations that consume this field (such as the Endpoint Picker) will treat an unset value as '0'.
-	// Criticality is used in flow control, primarily in the event of resource scarcity(reqeusts need to be queued).
-	// All requests will be queued, and flow control will _always_ allow requests of higher criticality to be served first.
-	// Fairness is only enforced and tracked between requests of the same criticality.
+	// Priority is used in flow control, primarily in the event of resource scarcity(reqeusts need to be queued).
+	// All requests will be queued, and flow control will _always_ allow requests of higher priority to be served first.
+	// Fairness is only enforced and tracked between requests of the same priority.
 	//
-	// Example: requests with Criticality 10 will always be served before
-	// requests with Criticality of 0(the value used if Criticality is unset or no InfereneceObjective is specified).
-	// Similarly requests with a Criticality of -10 will always be served after requests with Criticality of 0.
+	// Example: requests with Priority 10 will always be served before
+	// requests with Priority of 0 (the value used if Priority is unset or no InfereneceObjective is specified).
+	// Similarly requests with a Priority of -10 will always be served after requests with Priority of 0.
 	// +optional
-	Criticality *int `json:"criticality,omitempty"`
+	Priority *int `json:"priority,omitempty"`
 
 	// PoolRef is a reference to the inference pool, the pool must exist in the same namespace.
 	//

apix/v1alpha2/zz_generated.deepcopy.go

@@ -21,8 +21,8 @@ spec:
     - jsonPath: .spec.poolRef.name
       name: Inference Pool
       type: string
-    - jsonPath: .spec.criticality
-      name: Criticality
+    - jsonPath: .spec.priority
+      name: Priority
       type: string
     - jsonPath: .metadata.creationTimestamp
       name: Age
@@ -68,21 +68,6 @@ spec:
               creation timestamp, will be selected to remain valid. In the event of a race
               condition, one will be selected at random.
             properties:
-              criticality:
-                description: |-
-                  Criticality defines how important it is to serve the request compared to other requests in the same pool.
-                  Criticality is an integer value that defines the priority of the request.
-                  The higher the value, the more critical the request is; negative values _are_ allowed.
-                  No default value is set for this field, allowing for future additions of new fields that may 'one of' with this field.
-                  However, implementations that consume this field (such as the Endpoint Picker) will treat an unset value as '0'.
-                  Criticality is used in flow control, primarily in the event of resource scarcity(reqeusts need to be queued).
-                  All requests will be queued, and flow control will _always_ allow requests of higher criticality to be served first.
-                  Fairness is only enforced and tracked between requests of the same criticality.
-
-                  Example: requests with Criticality 10 will always be served before
-                  requests with Criticality of 0(the value used if Criticality is unset or no InfereneceObjective is specified).
-                  Similarly requests with a Criticality of -10 will always be served after requests with Criticality of 0.
-                type: integer
               poolRef:
                 description: PoolRef is a reference to the inference pool, the pool
                   must exist in the same namespace.
@@ -108,6 +93,21 @@ spec:
                 required:
                 - name
                 type: object
+              priority:
+                description: |-
+                  Priority defines how important it is to serve the request compared to other requests in the same pool.
+                  Priority is an integer value that defines the priority of the request.
+                  The higher the value, the more critical the request is; negative values _are_ allowed.
+                  No default value is set for this field, allowing for future additions of new fields that may 'one of' with this field.
+                  However, implementations that consume this field (such as the Endpoint Picker) will treat an unset value as '0'.
+                  Priority is used in flow control, primarily in the event of resource scarcity(reqeusts need to be queued).
+                  All requests will be queued, and flow control will _always_ allow requests of higher priority to be served first.
+                  Fairness is only enforced and tracked between requests of the same priority.
+
+                  Example: requests with Priority 10 will always be served before
+                  requests with Priority of 0 (the value used if Priority is unset or no InfereneceObjective is specified).
+                  Similarly requests with a Priority of -10 will always be served after requests with Priority of 0.
+                type: integer
             required:
             - poolRef
             type: object

client-go/applyconfiguration/apix/v1alpha2/inferenceobjectivespec.go

@@ -3,7 +3,7 @@ kind: InferenceObjective
 metadata:
   name: food-review
 spec:
-  criticality: 1
+  priority: 1
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -13,7 +13,7 @@ kind: InferenceObjective
 metadata:
   name: base-model
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -23,7 +23,7 @@ kind: InferenceObjective
 metadata:
   name: base-model-cpu
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct

config/crd/bases/inference.networking.x-k8s.io_inferenceobjectives.yaml

@@ -3,7 +3,7 @@ kind: InferenceObjective
 metadata:
   name: adapter-0
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -15,7 +15,7 @@ kind: InferenceObjective
 metadata:
   name: adapter-1
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -27,7 +27,7 @@ kind: InferenceObjective
 metadata:
   name: adapter-2
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -39,7 +39,7 @@ kind: InferenceObjective
 metadata:
   name: adapter-3
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -51,7 +51,7 @@ kind: InferenceObjective
 metadata:
   name: adapter-4
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -63,7 +63,7 @@ kind: InferenceObjective
 metadata:
   name: adapter-5
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -75,7 +75,7 @@ kind: InferenceObjective
 metadata:
   name: adapter-6
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -87,7 +87,7 @@ kind: InferenceObjective
 metadata:
   name: adapter-7
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -99,7 +99,7 @@ kind: InferenceObjective
 metadata:
   name: adapter-8
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -111,7 +111,7 @@ kind: InferenceObjective
 metadata:
   name: adapter-9
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -123,7 +123,7 @@ kind: InferenceObjective
 metadata:
   name: adapter-10
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -135,7 +135,7 @@ kind: InferenceObjective
 metadata:
   name: adapter-11
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -147,7 +147,7 @@ kind: InferenceObjective
 metadata:
   name: adapter-12
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -160,7 +160,7 @@ kind: InferenceObjective
 metadata:
   name: adapter-13
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -173,7 +173,7 @@ kind: InferenceObjective
 metadata:
   name: adapter-14
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
     name: vllm-llama3-8b-instruct
@@ -186,7 +186,7 @@ kind: InferenceObjective
 metadata:
   name: base-model
 spec:
-  criticality: 2
+  priority: 2
   poolRef:
     group: inference.networking.k8s.io
-    name: vllm-llama3-8b-instruct
+    name: vllm-llama3-8b-instruct

config/manifests/inferenceobjective.yaml

@@ -43,19 +43,19 @@ var (
 	pool          = utiltest.MakeInferencePool("test-pool1").Namespace("ns1").ObjRef()
 	infObjective1 = utiltest.MakeInferenceObjective("model1").
 			Namespace(pool.Namespace).
-			Criticality(1).
+			Priority(1).
 			CreationTimestamp(metav1.Unix(1000, 0)).
 			PoolName(pool.Name).
 			PoolGroup("inference.networking.k8s.io").ObjRef()
 	infObjective1Pool2 = utiltest.MakeInferenceObjective(infObjective1.Name).
 				Namespace(infObjective1.Namespace).
-				Criticality(*infObjective1.Spec.Criticality).
+				Priority(*infObjective1.Spec.Priority).
 				CreationTimestamp(metav1.Unix(1001, 0)).
 				PoolName("test-pool2").
 				PoolGroup("inference.networking.k8s.io").ObjRef()
 	infObjective1Critical = utiltest.MakeInferenceObjective(infObjective1.Name).
 				Namespace(infObjective1.Namespace).
-				Criticality(2).
+				Priority(2).
 				CreationTimestamp(metav1.Unix(1003, 0)).
 				PoolName(pool.Name).
 				PoolGroup("inference.networking.k8s.io").ObjRef()
@@ -67,7 +67,7 @@ var (
 				PoolGroup("inference.networking.k8s.io").ObjRef()
 	infObjective1DiffGroup = utiltest.MakeInferenceObjective(infObjective1.Name).
 				Namespace(pool.Namespace).
-				Criticality(1).
+				Priority(1).
 				CreationTimestamp(metav1.Unix(1005, 0)).
 				PoolName(pool.Name).
 				PoolGroup("inference.networking.x-k8s.io").ObjRef()
@@ -112,7 +112,7 @@ func TestInferenceObjectiveReconciler(t *testing.T) {
 			wantObjectives:     []*v1alpha2.InferenceObjective{},
 		},
 		{
-			name:               "Objective changed criticality",
+			name:               "Objective changed priority",
 			objectivessInStore: []*v1alpha2.InferenceObjective{infObjective1},
 			objective:          infObjective1Critical,
 			wantObjectives:     []*v1alpha2.InferenceObjective{infObjective1Critical},

config/manifests/regression-testing/inferenceobjective.yaml

@@ -114,7 +114,7 @@ func TestObjective(t *testing.T) {
 	model2ts := testutil.MakeInferenceObjective("model2").ObjRef()
 	// Same model name as model1ts, newer timestamp
 	model1tsCritical := testutil.MakeInferenceObjective("model1").
-		Criticality(2).ObjRef()
+		Priority(2).ObjRef()
 	// Same object name as model2ts, different model name.
 	model2chat := testutil.MakeInferenceObjective(model2ts.Name).ObjRef()
 
@@ -135,7 +135,7 @@ func TestObjective(t *testing.T) {
 			wantOpResult: true,
 		},
 		{
-			name:           "Set model1 with the same modelName, but with diff criticality, should update.",
+			name:           "Set model1 with the same modelName, but with diff priority, should update.",
 			existingModels: []*v1alpha2.InferenceObjective{model1ts},
 			op: func(ds Datastore) bool {
 				ds.ObjectiveSet(model1tsCritical)