[target-allocator] Introduce "per node" allocation strategy to target allocator

.chloggen/per-node-allocation-strategy.yaml

@@ -0,0 +1,16 @@
+# One of 'breaking', 'deprecation', 'new_component', 'enhancement', 'bug_fix'
+change_type: enhancement
+
+# The name of the component, or a single word describing the area of concern, (e.g. operator, target allocator, github action)
+component: target allocator
+
+# A brief description of the change. Surround your text with quotes ("") if it needs to start with a backtick (`).
+note: Add new "per node" allocation strategy to target allocator. This strategy will allocate targets to nodes on which given target resides.  It should only be used conjunction with the daemonset mode.
+
+# One or more tracking issues related to the change
+issues: [1828]
+
+# (Optional) One or more lines of additional information to render under the primary note.
+# These lines will be padded with 2 spaces and then inserted directly into the document.
+# Use pipe (|) for multiline entries.
+subtext:

.github/workflows/e2e.yaml

@@ -30,6 +30,7 @@ jobs:
          - e2e-autoscale
          - e2e-pdb
          - e2e-opampbridge
+         - e2e-targetallocator
          - e2e-prometheuscr
          - e2e-multi-instrumentation
        include:

Makefile

@@ -246,6 +246,11 @@ e2e-multi-instrumentation:
 e2e-opampbridge:
 	$(KUTTL) test --config kuttl-test-opampbridge.yaml
 
+# Target allocator end-to-tests
+.PHONY: e2e-targetallocator
+e2e-targetallocator:
+	$(KUTTL) test --config kuttl-test-targetallocator.yaml
+
 .PHONY: prepare-e2e
 prepare-e2e: kuttl set-image-controller add-image-targetallocator add-image-opampbridge container container-target-allocator container-operator-opamp-bridge start-kind cert-manager install-metrics-server install-targetallocator-prometheus-crds load-image-all deploy
 

apis/v1alpha1/allocation_strategy.go

@@ -16,7 +16,7 @@ package v1alpha1
 
 type (
 	// OpenTelemetryTargetAllocatorAllocationStrategy represent which strategy to distribute target to each collector
-	// +kubebuilder:validation:Enum=least-weighted;consistent-hashing
+	// +kubebuilder:validation:Enum=least-weighted;consistent-hashing;per-node
 	OpenTelemetryTargetAllocatorAllocationStrategy string
 )
 
@@ -26,4 +26,7 @@ const (
 
 	// OpenTelemetryTargetAllocatorAllocationStrategyConsistentHashing targets will be consistently added to collectors, which allows a high-availability setup.
 	OpenTelemetryTargetAllocatorAllocationStrategyConsistentHashing OpenTelemetryTargetAllocatorAllocationStrategy = "consistent-hashing"
+
+	// OpenTelemetryTargetAllocatorAllocationStrategyPerNode targets will be assigned to the collector on the node they reside on (use only with daemon set).
+	OpenTelemetryTargetAllocatorAllocationStrategyPerNode OpenTelemetryTargetAllocatorAllocationStrategy = "per-node"
 )

apis/v1alpha1/collector_webhook.go

@@ -228,32 +228,14 @@ func (c CollectorWebhook) validate(ctx context.Context, r *OpenTelemetryCollecto
 		return warnings, fmt.Errorf("the OpenTelemetry Collector mode is set to %s, which does not support the attribute 'AdditionalContainers'", r.Spec.Mode)
 	}
 
-	// validate target allocation
-	if r.Spec.TargetAllocator.Enabled && r.Spec.Mode != ModeStatefulSet {
-		return warnings, fmt.Errorf("the OpenTelemetry Collector mode is set to %s, which does not support the target allocation deployment", r.Spec.Mode)
-	}
-
-	// validate Prometheus config for target allocation
+	// validate target allocator configs
 	if r.Spec.TargetAllocator.Enabled {
-		promCfg, err := ta.ConfigToPromConfig(r.Spec.Config)
-		if err != nil {
-			return warnings, fmt.Errorf("the OpenTelemetry Spec Prometheus configuration is incorrect, %w", err)
-		}
-		err = ta.ValidatePromConfig(promCfg, r.Spec.TargetAllocator.Enabled, featuregate.EnableTargetAllocatorRewrite.IsEnabled())
+		taWarnings, err := c.validateTargetAllocatorConfig(ctx, r)
 		if err != nil {
-			return warnings, fmt.Errorf("the OpenTelemetry Spec Prometheus configuration is incorrect, %w", err)
+			return warnings, err
 		}
-		err = ta.ValidateTargetAllocatorConfig(r.Spec.TargetAllocator.PrometheusCR.Enabled, promCfg)
-		if err != nil {
-			return warnings, fmt.Errorf("the OpenTelemetry Spec Prometheus configuration is incorrect, %w", err)
-		}
-		// if the prometheusCR is enabled, it needs a suite of permissions to function
-		if r.Spec.TargetAllocator.PrometheusCR.Enabled {
-			if subjectAccessReviews, err := c.reviewer.CheckPolicyRules(ctx, r.GetNamespace(), r.Spec.TargetAllocator.ServiceAccount, targetAllocatorCRPolicyRules...); err != nil {
-				return warnings, fmt.Errorf("unable to check rbac rules %w", err)
-			} else if allowed, deniedReviews := rbac.AllSubjectAccessReviewsAllowed(subjectAccessReviews); !allowed {
-				warnings = append(warnings, warningsGroupedByResource(deniedReviews)...)
-			}
+		if taWarnings != nil {
+			warnings = append(warnings, taWarnings...)
 		}
 	}
 
@@ -360,6 +342,44 @@ func (c CollectorWebhook) validate(ctx context.Context, r *OpenTelemetryCollecto
 	return warnings, nil
 }
 
+func (c CollectorWebhook) validateTargetAllocatorConfig(ctx context.Context, r *OpenTelemetryCollector) (admission.Warnings, error) {
+	if r.Spec.Mode != ModeStatefulSet && r.Spec.Mode != ModeDaemonSet {
+		return nil, fmt.Errorf("the OpenTelemetry Collector mode is set to %s, which does not support the target allocation deployment", r.Spec.Mode)
+	}
+
+	if r.Spec.Mode == ModeDaemonSet && r.Spec.TargetAllocator.AllocationStrategy != OpenTelemetryTargetAllocatorAllocationStrategyPerNode {
+		return nil, fmt.Errorf("the OpenTelemetry Collector mode is set to %s, which must be used with target allocation strategy %s ", r.Spec.Mode, OpenTelemetryTargetAllocatorAllocationStrategyPerNode)
+	}
+
+	if r.Spec.TargetAllocator.AllocationStrategy == OpenTelemetryTargetAllocatorAllocationStrategyPerNode && r.Spec.Mode != ModeDaemonSet {
+		return nil, fmt.Errorf("target allocation strategy %s is only supported in OpenTelemetry Collector mode %s", OpenTelemetryTargetAllocatorAllocationStrategyPerNode, ModeDaemonSet)
+	}
+
+	// validate Prometheus config for target allocation
+	promCfg, err := ta.ConfigToPromConfig(r.Spec.Config)
+	if err != nil {
+		return nil, fmt.Errorf("the OpenTelemetry Spec Prometheus configuration is incorrect, %w", err)
+	}
+	err = ta.ValidatePromConfig(promCfg, r.Spec.TargetAllocator.Enabled, featuregate.EnableTargetAllocatorRewrite.IsEnabled())
+	if err != nil {
+		return nil, fmt.Errorf("the OpenTelemetry Spec Prometheus configuration is incorrect, %w", err)
+	}
+	err = ta.ValidateTargetAllocatorConfig(r.Spec.TargetAllocator.PrometheusCR.Enabled, promCfg)
+	if err != nil {
+		return nil, fmt.Errorf("the OpenTelemetry Spec Prometheus configuration is incorrect, %w", err)
+	}
+	// if the prometheusCR is enabled, it needs a suite of permissions to function
+	if r.Spec.TargetAllocator.PrometheusCR.Enabled {
+		if subjectAccessReviews, err := c.reviewer.CheckPolicyRules(ctx, r.GetNamespace(), r.Spec.TargetAllocator.ServiceAccount, targetAllocatorCRPolicyRules...); err != nil {
+			return nil, fmt.Errorf("unable to check rbac rules %w", err)
+		} else if allowed, deniedReviews := rbac.AllSubjectAccessReviewsAllowed(subjectAccessReviews); !allowed {
+			return warningsGroupedByResource(deniedReviews), nil
+		}
+	}
+
+	return nil, nil
+}
+
 func checkAutoscalerSpec(autoscaler *AutoscalerSpec) error {
 	if autoscaler.Behavior != nil {
 		if autoscaler.Behavior.ScaleDown != nil && autoscaler.Behavior.ScaleDown.StabilizationWindowSeconds != nil &&

apis/v1alpha1/collector_webhook_test.go

@@ -669,6 +669,19 @@ func TestOTELColValidatingWebhook(t *testing.T) {
 			},
 			expectedErr: "the OpenTelemetry Spec Prometheus configuration is incorrect",
 		},
+		{
+			name: "invalid target allocation strategy",
+			otelcol: OpenTelemetryCollector{
+				Spec: OpenTelemetryCollectorSpec{
+					Mode: ModeDaemonSet,
+					TargetAllocator: OpenTelemetryTargetAllocator{
+						Enabled:            true,
+						AllocationStrategy: OpenTelemetryTargetAllocatorAllocationStrategyLeastWeighted,
+					},
+				},
+			},
+			expectedErr: "mode is set to daemonset, which must be used with target allocation strategy per-node",
+		},
 		{
 			name: "invalid port name",
 			otelcol: OpenTelemetryCollector{

apis/v1alpha1/opentelemetrycollector_types.go

@@ -297,7 +297,8 @@ type OpenTelemetryTargetAllocator struct {
 	// +optional
 	Resources v1.ResourceRequirements `json:"resources,omitempty"`
 	// AllocationStrategy determines which strategy the target allocator should use for allocation.
-	// The current options are least-weighted and consistent-hashing. The default option is consistent-hashing
+	// The current options are least-weighted, consistent-hashing and per-node. The default is
+	// consistent-hashing.
 	// +optional
 	// +kubebuilder:default:=consistent-hashing
 	AllocationStrategy OpenTelemetryTargetAllocatorAllocationStrategy `json:"allocationStrategy,omitempty"`

bundle/manifests/opentelemetry.io_opentelemetrycollectors.yaml

@@ -5027,11 +5027,12 @@ spec:
                     default: consistent-hashing
                     description: AllocationStrategy determines which strategy the
                       target allocator should use for allocation. The current options
-                      are least-weighted and consistent-hashing. The default option
-                      is consistent-hashing
+                      are least-weighted, consistent-hashing and per-node. The default
+                      is consistent-hashing.
                     enum:
                     - least-weighted
                     - consistent-hashing
+                    - per-node
                     type: string
                   enabled:
                     description: Enabled indicates whether to use a target allocation

cmd/otel-allocator/allocation/allocatortest.go

@@ -51,6 +51,7 @@ func MakeNCollectors(n int, startingIndex int) map[string]*Collector {
 		toReturn[collector] = &Collector{
 			Name:       collector,
 			NumTargets: 0,
+			NodeName:   fmt.Sprintf("node-%d", i),
 		}
 	}
 	return toReturn

cmd/otel-allocator/allocation/consistent_hashing.go

@@ -161,7 +161,7 @@ func (c *consistentHashingAllocator) handleCollectors(diff diff.Changes[*Collect
 	}
 	// Insert the new collectors
 	for _, i := range diff.Additions() {
-		c.collectors[i.Name] = NewCollector(i.Name)
+		c.collectors[i.Name] = NewCollector(i.Name, i.NodeName)
 		c.consistentHasher.Add(c.collectors[i.Name])
 	}
 

cmd/otel-allocator/allocation/least_weighted.go

@@ -191,7 +191,7 @@ func (allocator *leastWeightedAllocator) handleCollectors(diff diff.Changes[*Col
 	}
 	// Insert the new collectors
 	for _, i := range diff.Additions() {
-		allocator.collectors[i.Name] = NewCollector(i.Name)
+		allocator.collectors[i.Name] = NewCollector(i.Name, i.NodeName)
 	}
 	if allocateTargets {
 		for _, item := range allocator.targetItems {