Merge pull request #119 from dnorman3/dnorman3/priority-system

dnorman3/priority system

Author: dnorman3

Makefile

@@ -1,4 +1,4 @@
-VERSION = 1.10.1
+VERSION = 1.10.2
 # IMPORTANT! Update api version if a new release affects cnr
 API_VERSION = 1.0.0
 IMAGE = cyclops:$(VERSION)

README.md

@@ -35,8 +35,8 @@ See [Docs](docs/README.md)
 
 ## Requirements
 
-- [Kubernetes](https://kubernetes.io/) version 1.13+. Cyclops has been tested and deployed on 1.13+ and newer. Older versions of Kubernetes may have bugs or issues that will prevent it from functioning properly.
-- [Go](https://golang.org/) version 1.13+. Newer versions of Go are highly recommended.
+- [Kubernetes](https://kubernetes.io/) version 1.32+. Cyclops has been tested and deployed on 1.32+ and newer. Older versions of Kubernetes may have bugs or issues that will prevent it from functioning properly.
+- [Go](https://golang.org/) version 1.23+. Newer versions of Go are highly recommended.
 - Dependencies and their locked versions can be found in `go.mod` and `go.sum`.
 
 ## Building

deploy/crds/atlassian.com_nodegroups_crd.yaml

@@ -306,6 +306,12 @@ spec:
                   - triggerEndpoint
                   type: object
                 type: array
+              priority:
+                description: Priority controls the ordering of CNR creation for this
+                  NodeGroup. Lower values are higher priority. -10 runs before 0;
+                  then 10, 20, ...  Value range is from -2,147,483,648 to 2,147,483,647.
+                format: int32
+                type: integer
               skipInitialHealthChecks:
                 description: SkipInitialHealthChecks is an optional flag to skip the
                   initial set of node health checks before cycling begins This does

pkg/apis/atlassian/v1/nodegroup_types.go

@@ -39,6 +39,10 @@ type NodeGroupSpec struct {
 
 	// SkipPreTerminationChecks is an optional flag to skip pre-termination checks during cycling
 	SkipPreTerminationChecks bool `json:"skipPreTerminationChecks,omitempty"`
+
+	// Priority controls the ordering of CNR creation for this NodeGroup.
+	// Lower values are higher priority. Examples: -10 runs before 0; then 10, 20, ...
+    Priority int32 `json:"priority,omitempty"`
 }
 
 // NodeGroupStatus defines the observed state of NodeGroup

pkg/metrics/metrics.go

@@ -3,6 +3,7 @@ package metrics
 import (
 	"context"
 	"fmt"
+	"strings"
 	"time"
 
 	v1 "github.com/atlassian-labs/cyclops/pkg/apis/atlassian/v1"
@@ -43,6 +44,24 @@ var (
 		[]string{"phase"},
 		nil,
 	)
+	// NodeGroupInfo provides static information about nodegroups
+	NodeGroupInfo = prometheus.NewDesc(
+		fmt.Sprintf("%v_node_group_info", namespace),
+		"Static information about nodegroups in the cluster",
+		[]string{
+			"nodegroup_name",
+			"nodegroups_list",
+			"node_selector",           
+			"concurrency",
+			"method",
+			"max_failed_cnrs",
+			"skip_missing_named_nodes",
+			"skip_initial_health_checks",
+			"skip_pre_termination_checks",
+			"priority",
+		},
+		nil,
+	)
 )
 
 // Register registers the custom metrics with prometheus
@@ -105,6 +124,7 @@ func (c cyclopsCollector) Describe(ch chan<- *prometheus.Desc) {
 	ch <- CycleNodeRequestsByPhase
 	ch <- CycleNodeStatuses
 	ch <- CycleNodeStatusesByPhase
+	ch <- NodeGroupInfo
 }
 
 func (c cyclopsCollector) Collect(ch chan<- prometheus.Metric) {
@@ -172,4 +192,22 @@ func (c cyclopsCollector) Collect(ch chan<- prometheus.Metric) {
 		prometheus.GaugeValue,
 		float64(len(c.cycleNodeStatusList.Items)),
 	)
+
+    for _, nodegroup := range c.nodeGroupList.Items {
+        ch <- prometheus.MustNewConstMetric(
+            NodeGroupInfo,
+            prometheus.GaugeValue,
+            1.0,
+			nodegroup.Spec.NodeGroupName,
+			strings.Join(nodegroup.Spec.NodeGroupsList, ","),
+			nodegroup.Spec.NodeSelector.String(),
+			fmt.Sprintf("%d", nodegroup.Spec.CycleSettings.Concurrency),
+			string(nodegroup.Spec.CycleSettings.Method),
+			fmt.Sprintf("%d", nodegroup.Spec.MaxFailedCycleNodeRequests),
+			fmt.Sprintf("%t", nodegroup.Spec.ValidationOptions.SkipMissingNamedNodes),
+			fmt.Sprintf("%t", nodegroup.Spec.SkipInitialHealthChecks),
+			fmt.Sprintf("%t", nodegroup.Spec.SkipPreTerminationChecks),
+			fmt.Sprintf("%d", nodegroup.Spec.Priority),
+        )
+    }
 }

pkg/metrics/metrics_test.go

@@ -0,0 +1,128 @@
+package metrics
+
+import (
+	"testing"
+
+	v1 "github.com/atlassian-labs/cyclops/pkg/apis/atlassian/v1"
+	"github.com/prometheus/client_golang/prometheus"
+	"github.com/stretchr/testify/assert"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+)
+
+func TestNodeGroupInfoMetric(t *testing.T) {
+	// Create a test NodeGroup
+	testNodeGroup := &v1.NodeGroup{
+		ObjectMeta: metav1.ObjectMeta{
+			Name: "test-ng",
+		},
+		Spec: v1.NodeGroupSpec{
+			NodeGroupName: "test-nodegroup",
+			NodeGroupsList: []string{"ng1", "ng2"},
+			NodeSelector: metav1.LabelSelector{
+				MatchLabels: map[string]string{
+					"role": "worker",
+				},
+			},
+			CycleSettings: v1.CycleSettings{
+				Method:      "Drain",
+				Concurrency: 2,
+			},
+			MaxFailedCycleNodeRequests: 3,
+			ValidationOptions: v1.ValidationOptions{
+				SkipMissingNamedNodes: true,
+			},
+			SkipInitialHealthChecks: false,
+			SkipPreTerminationChecks: true,
+			Priority: 10,
+		},
+	}
+
+	// Create a test NodeGroupList
+	testNodeGroupList := &v1.NodeGroupList{
+		Items: []v1.NodeGroup{*testNodeGroup},
+	}
+
+	// Create a mock collector with all required fields initialized
+	collector := &cyclopsCollector{
+		nodeGroupList:        testNodeGroupList,
+		cycleNodeRequestList: &v1.CycleNodeRequestList{Items: []v1.CycleNodeRequest{}},
+		cycleNodeStatusList:  &v1.CycleNodeStatusList{Items: []v1.CycleNodeStatus{}},
+	}
+
+	// Test that the metric can be described
+	t.Run("Describe", func(t *testing.T) {
+		ch := make(chan *prometheus.Desc, 10) // Buffer size to hold all metrics
+		go func() {
+			collector.Describe(ch)
+			close(ch)
+		}()
+		
+		// Read all descriptions
+		descriptions := make([]*prometheus.Desc, 0)
+		for desc := range ch {
+			descriptions = append(descriptions, desc)
+		}
+		
+		// Verify we got descriptions and one contains our metric
+		assert.Greater(t, len(descriptions), 0)
+		found := false
+		for _, desc := range descriptions {
+			if desc.String() == NodeGroupInfo.String() {
+				found = true
+				break
+			}
+		}
+		assert.True(t, found, "NodeGroupInfo metric should be described")
+	})
+
+	// Test that the metric can be collected
+	t.Run("Collect", func(t *testing.T) {
+		ch := make(chan prometheus.Metric, 10) // Buffer size to hold all metrics
+		go func() {
+			collector.Collect(ch)
+			close(ch)
+		}()
+		
+		// Read all metrics
+		metrics := make([]prometheus.Metric, 0)
+		for metric := range ch {
+			metrics = append(metrics, metric)
+		}
+		
+		// Verify we got metrics and one is our NodeGroupInfo
+		assert.Greater(t, len(metrics), 0)
+		found := false
+		for _, metric := range metrics {
+			desc := metric.Desc()
+			if desc.String() == NodeGroupInfo.String() {
+				found = true
+				break
+			}
+		}
+		assert.True(t, found, "NodeGroupInfo metric should be collected")
+	})
+}
+
+func TestNodeGroupInfoLabels(t *testing.T) {
+	// Test that the metric definition has the correct number of labels
+	expectedLabels := []string{
+		"nodegroup_name",
+		"nodegroups_list", 
+		"node_selector",
+		"concurrency",
+		"method",
+		"max_failed_cnrs",
+		"skip_missing_named_nodes",
+		"skip_initial_health_checks",
+		"skip_pre_termination_checks",
+		"priority",
+	}
+	
+	// Verify we have the expected number of labels
+	assert.Equal(t, 10, len(expectedLabels))
+	
+	assert.Contains(t, NodeGroupInfo.String(), "cyclops_node_group_info")
+	
+	// Test that the metric is properly registered
+	assert.NotNil(t, NodeGroupInfo)
+}