feat(controller): Track N/M progress. See argoproj#2717 (argoproj#4194)

Signed-off-by: Alex Capras <alexcapras@gmail.com>

api/openapi-spec/swagger.json

@@ -3063,6 +3063,10 @@
           "description": "PodIP captures the IP of the pod for daemoned steps",
           "type": "string"
         },
+        "progress": {
+          "description": "Progress to completion",
+          "type": "string"
+        },
         "resourcesDuration": {
           "description": "ResourcesDuration is indicative, but not accurate, resource duration. This is populated when the nodes completes.",
           "type": "object",
@@ -4627,6 +4631,10 @@
           "description": "Phase a simple, high-level summary of where the workflow is in its lifecycle.",
           "type": "string"
         },
+        "progress": {
+          "description": "Progress to completion",
+          "type": "string"
+        },
         "resourcesDuration": {
           "description": "ResourcesDuration is the total for the workflow",
           "type": "object",

cmd/argo/commands/get.go

@@ -145,6 +145,7 @@ func printWorkflowHelper(wf *wfv1.Workflow, getArgs getFlags) string {
 		if wf.Status.EstimatedDuration > 0 {
 			out += fmt.Sprintf(fmtStr, "EstimatedDuration:", humanize.Duration(wf.Status.EstimatedDuration.ToDuration()))
 		}
+		out += fmt.Sprintf(fmtStr, "Progress:", wf.Status.Progress)
 	}
 	if !wf.Status.ResourcesDuration.IsZero() {
 		out += fmt.Sprintf(fmtStr, "ResourcesDuration:", wf.Status.ResourcesDuration)

cmd/argo/commands/get_test.go

@@ -109,6 +109,16 @@ func TestStatusToNodeFieldSelector(t *testing.T) {
 }
 
 func Test_printWorkflowHelper(t *testing.T) {
+	t.Run("Progress", func(t *testing.T) {
+		var wf wfv1.Workflow
+		testutil.MustUnmarshallYAML(`
+status:
+  phase: Running
+  progress: 1/2
+`, &wf)
+		output := printWorkflowHelper(&wf, getFlags{})
+		assert.Regexp(t, `Progress: *1/2`, output)
+	})
 	t.Run("EstimatedDuration", func(t *testing.T) {
 		var wf wfv1.Workflow
 		testutil.MustUnmarshallYAML(`

docs/fields.md

@@ -672,6 +672,7 @@ WorkflowStatus contains overall status information about a workflow
 |`outputs`|[`Outputs`](#outputs)|Outputs captures output values and artifact locations produced by the workflow via global outputs|
 |`persistentVolumeClaims`|`Array<`[`Volume`](#volume)`>`|PersistentVolumeClaims tracks all PVCs that were created as part of the io.argoproj.workflow.v1alpha1. The contents of this list are drained at the end of the workflow.|
 |`phase`|`string`|Phase a simple, high-level summary of where the workflow is in its lifecycle.|
+|`progress`|`string`|Progress to completion|
 |`resourcesDuration`|`Map< integer , int64 >`|ResourcesDuration is the total for the workflow|
 |`startedAt`|[`Time`](#time)|Time at which this workflow started|
 |`storedTemplates`|[`Template`](#template)|StoredTemplates is a mapping between a template ref and the node's status.|
@@ -1996,6 +1997,7 @@ NodeStatus contains status information about an individual node in the workflow
 |`outputs`|[`Outputs`](#outputs)|Outputs captures output parameter values and artifact locations produced by this template invocation|
 |`phase`|`string`|Phase a simple, high-level summary of where the node is in its lifecycle. Can be used as a state machine.|
 |`podIP`|`string`|PodIP captures the IP of the pod for daemoned steps|
+|`progress`|`string`|Progress to completion|
 |`resourcesDuration`|`Map< integer , int64 >`|ResourcesDuration is indicative, but not accurate, resource duration. This is populated when the nodes completes.|
 |`startedAt`|[`Time`](#time)|Time at which this node started|
 |~`storedTemplateID`~|~`string`~|~StoredTemplateID is the ID of stored template.~ DEPRECATED: This value is not used anymore.|

docs/progress.md

@@ -0,0 +1,26 @@
+# Workflow Progress
+
+![alpha](assets/alpha.svg)
+
+> v2.12 and after
+
+When you run a workflow, the controller will report on its progress.
+
+We define progress as two numbers, `N/M` such that `0 <= N <= M and 0 <= M <= 1`. 
+
+* `N` is the number of completed tasks.
+* `M` is the total number of tasks.
+
+E.g. `0/0`, `0/1` or `50/100`.
+
+Unlike [estimated duration](estimated-duration.md), progress is deterministic. I.e. it will be the same for each workflow, regardless of any problems. 
+
+Progress for each node is calculated as follows:
+
+2. For a pod node either `1/1` if completed or `0/1` otherwise.
+3. For non-leaf nodes, the sum of its children.
+
+For a whole workflow's, progress is the sum of all its leaf nodes.
+
+!!! Warning 
+    `M` will increase during workflow run each time a node is added to the graph.

manifests/base/crds/full/argoproj.io_workflows.yaml

@@ -7502,6 +7502,8 @@ spec:
                     type: string
                   podIP:
                     type: string
+                  progress:
+                    type: string
                   resourcesDuration:
                     additionalProperties:
                       format: int64
@@ -8490,6 +8492,8 @@ spec:
               type: array
             phase:
               type: string
+            progress:
+              type: string
             resourcesDuration:
               additionalProperties:
                 format: int64

mkdocs.yml

@@ -42,6 +42,7 @@ nav:
           - artifact-repository-ref.md
           - resource-duration.md
           - estimated-duration.md
+          - progress.md
           - workflow-creator.md
           - synchronization.md
           - workflow-of-workflows.md

pkg/apis/workflow/v1alpha1/progress.go

@@ -0,0 +1,44 @@
+package v1alpha1
+
+import (
+	"fmt"
+	"strconv"
+	"strings"
+)
+
+// Progress in N/M format. N is number of task complete. M is number of tasks.
+type Progress string
+
+func NewProgress(n, m int64) (Progress, bool) {
+	return ParseProgress(fmt.Sprintf("%v/%v", n, m))
+}
+
+func ParseProgress(s string) (Progress, bool) {
+	v := Progress(s)
+	return v, v.IsValid()
+}
+
+func (in Progress) parts() []string {
+	return strings.SplitN(string(in), "/", 2)
+}
+
+func (in Progress) N() int64 {
+	return parseInt64(in.parts()[0])
+}
+
+func (in Progress) M() int64 {
+	return parseInt64(in.parts()[1])
+}
+
+func (in Progress) Add(x Progress) Progress {
+	return Progress(fmt.Sprintf("%v/%v", in.N()+x.N(), in.M()+x.M()))
+}
+
+func (in Progress) IsValid() bool {
+	return in != "" && in.N() >= 0 && in.N() <= in.M() && in.M() > 0
+}
+
+func parseInt64(s string) int64 {
+	v, _ := strconv.ParseInt(s, 10, 64)
+	return v
+}

pkg/apis/workflow/v1alpha1/progress_test.go

@@ -0,0 +1,28 @@
+package v1alpha1
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestProgress(t *testing.T) {
+	t.Run("ParseProgress", func(t *testing.T) {
+		_, ok := ParseProgress("")
+		assert.False(t, ok)
+		progress, ok := ParseProgress("0/1")
+		assert.True(t, ok)
+		assert.Equal(t, Progress("0/1"), progress)
+	})
+	t.Run("IsValid", func(t *testing.T) {
+		assert.False(t, Progress("").IsValid())
+		assert.False(t, Progress("/0").IsValid())
+		assert.False(t, Progress("0/").IsValid())
+		assert.False(t, Progress("0/0").IsValid())
+		assert.False(t, Progress("1/0").IsValid())
+		assert.True(t, Progress("0/1").IsValid())
+	})
+	t.Run("Add", func(t *testing.T) {
+		assert.Equal(t, Progress("1/2"), Progress("0/0").Add("1/2"))
+	})
+}

pkg/apis/workflow/v1alpha1/workflow_types.go

@@ -68,15 +68,6 @@ const (
 	NodeTypeSuspend   NodeType = "Suspend"
 )
 
-func (t NodeType) IsLeaf() bool {
-	switch t {
-	case NodeTypePod, NodeTypeRetry, NodeTypeSkipped, NodeTypeSuspend:
-		return true
-	default:
-		return false
-	}
-}
-
 // PodGCStrategy is the strategy when to delete completed pods for GC.
 type PodGCStrategy string
 
@@ -1127,6 +1118,9 @@ type WorkflowStatus struct {
 	// EstimatedDuration in seconds.
 	EstimatedDuration EstimatedDuration `json:"estimatedDuration,omitempty" protobuf:"varint,16,opt,name=estimatedDuration,casttype=EstimatedDuration"`
 
+	// Progress to completion
+	Progress Progress `json:"progress,omitempty" protobuf:"bytes,17,opt,name=progress,casttype=Progress"`
+
 	// A human readable message indicating details about why the workflow is in this condition.
 	Message string `json:"message,omitempty" protobuf:"bytes,4,opt,name=message"`
 
@@ -1391,6 +1385,9 @@ type NodeStatus struct {
 	// EstimatedDuration in seconds.
 	EstimatedDuration EstimatedDuration `json:"estimatedDuration,omitempty" protobuf:"varint,24,opt,name=estimatedDuration,casttype=EstimatedDuration"`
 
+	// Progress to completion
+	Progress Progress `json:"progress,omitempty" protobuf:"bytes,26,opt,name=progress,casttype=Progress"`
+
 	// ResourcesDuration is indicative, but not accurate, resource duration. This is populated when the nodes completes.
 	ResourcesDuration ResourcesDuration `json:"resourcesDuration,omitempty" protobuf:"bytes,21,opt,name=resourcesDuration"`
 
@@ -1566,10 +1563,6 @@ func (n NodeStatus) GetDuration() time.Duration {
 	return n.FinishedAt.Sub(n.StartedAt.Time)
 }
 
-func (in *NodeStatus) IsLeaf() bool {
-	return in.Type.IsLeaf()
-}
-
 // S3Bucket contains the access information required for interfacing with an S3 bucket
 type S3Bucket struct {
 	// Endpoint is the hostname of the bucket endpoint

test/e2e/progress_test.go

@@ -0,0 +1,36 @@
+// +build e2e
+
+package e2e
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/suite"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+
+	wfv1 "github.com/argoproj/argo/pkg/apis/workflow/v1alpha1"
+	"github.com/argoproj/argo/test/e2e/fixtures"
+)
+
+type ProgressSuite struct {
+	fixtures.E2ESuite
+}
+
+func (s *ProgressSuite) TestDefaultProgress() {
+	s.Given().
+		Workflow("@testdata/basic-workflow.yaml").
+		When().
+		SubmitWorkflow().
+		WaitForWorkflow().
+		Then().
+		ExpectWorkflow(func(t *testing.T, metadata *metav1.ObjectMeta, status *wfv1.WorkflowStatus) {
+			assert.Equal(t, wfv1.NodeSucceeded, status.Phase)
+			assert.Equal(t, wfv1.Progress("1/1"), status.Progress)
+			assert.Equal(t, wfv1.Progress("1/1"), status.Nodes[metadata.Name].Progress)
+		})
+}
+
+func TestProgressSuite(t *testing.T) {
+	suite.Run(t, new(ProgressSuite))
+}

ui/src/app/shared/services/workflows-service.ts

@@ -33,6 +33,7 @@ export class WorkflowsService {
             'items.status.finishedAt',
             'items.status.startedAt',
             'items.status.estimatedDuration',
+            'items.status.progress',
             'items.spec.suspend'
         ]
     ) {
@@ -83,6 +84,7 @@ export class WorkflowsService {
             'result.object.status.phase',
             'result.object.status.startedAt',
             'result.object.status.estimatedDuration',
+            'result.object.status.progress',
             'result.type',
             'result.object.metadata.labels',
             'result.object.spec.suspend'

ui/src/app/workflows/components/workflow-node-info/workflow-node-info.tsx

@@ -64,6 +64,7 @@ export const WorkflowNodeSummary = (props: Props) => {
             title: 'DURATION',
             value: <Ticker>{now => <DurationPanel duration={nodeDuration(props.node, now)} phase={props.node.phase} estimatedDuration={props.node.estimatedDuration} />}</Ticker>
         },
+        {title: 'PROGRESS', value: props.node.progress || '-'},
         {
             title: 'MEMOIZATION',
             value: (

ui/src/app/workflows/components/workflow-summary-panel.tsx

@@ -46,7 +46,8 @@ export const WorkflowSummaryPanel = (props: {workflow: Workflow}) => (
                             estimatedDuration={props.workflow.status.estimatedDuration}
                         />
                     )
-                }
+                },
+                {title: 'Progress', value: props.workflow.status.progress || '-'}
             ];
             const creator = props.workflow.metadata.labels[labels.creator];
             if (creator) {

ui/src/app/workflows/components/workflows-list/workflows-list.tsx

@@ -314,6 +314,7 @@ export class WorkflowsList extends BasePage<RouteComponentProps<any>, State> {
                             <div className='columns small-2'>STARTED</div>
                             <div className='columns small-2'>FINISHED</div>
                             <div className='columns small-1'>DURATION</div>
+                            <div className='columns small-1'>PROGRESS</div>
                             <div className='columns small-1'>DETAILS</div>
                         </div>
                     </div>

ui/src/app/workflows/components/workflows-row/workflows-row.tsx

@@ -56,9 +56,10 @@ export class WorkflowsRow extends React.Component<WorkflowsRowProps, WorkflowRow
                         <div className='columns small-2'>
                             <Timestamp date={wf.status.finishedAt} />
                         </div>
-                        <div className='columns small-2'>
+                        <div className='columns small-1'>
                             <Ticker>{() => <DurationPanel phase={wf.status.phase} duration={wfDuration(wf.status)} estimatedDuration={wf.status.estimatedDuration} />}</Ticker>
                         </div>
+                        <div className='columns small-1'>{wf.status.progress || '-'}</div>
                         <div className='columns small-1'>
                             <div className='workflows-list__labels-container'>
                                 <div

ui/src/models/workflows.ts

@@ -459,6 +459,11 @@ export interface NodeStatus {
      */
     estimatedDuration?: number;
 
+    /**
+     * Progress as numerator/denominator.
+     */
+    progress?: string;
+
     /**
      * How much resource was requested.
      */
@@ -564,6 +569,10 @@ export interface WorkflowStatus {
      */
     estimatedDuration?: number;
 
+    /**
+     * Progress as numerator/denominator.
+     */
+    progress?: string;
     /**
      * A human readable message indicating details about why the workflow is in this condition.
      */