Expire stuck plugin jobs (#8492)

* Add stale job expiry and expire API

* Add expire job button

* Add test hook and coverage for ExpirePluginJobAPI

* Document scheduler filtering side effect and reuse helper

* Restore job spec proposal test

* Regenerate plugin template output

---------

Co-authored-by: Copilot <copilot@github.com>

weed/admin/dash/admin_server.go

@@ -98,6 +98,7 @@ type AdminServer struct {
 	// Maintenance system
 	maintenanceManager *maintenance.MaintenanceManager
 	plugin             *adminplugin.Plugin
+	expireJobHandler   func(jobID string, reason string) (*adminplugin.TrackedJob, bool, error)
 
 	// Topic retention purger
 	topicRetentionPurger *TopicRetentionPurger
@@ -1020,6 +1021,17 @@ func (s *AdminServer) GetPluginJobDetail(jobID string, activityLimit, relatedLim
 	return s.plugin.BuildJobDetail(jobID, activityLimit, relatedLimit)
 }
 
+// ExpirePluginJob marks an active plugin job as failed so it no longer blocks scheduling.
+func (s *AdminServer) ExpirePluginJob(jobID, reason string) (*adminplugin.TrackedJob, bool, error) {
+	if handler := s.expireJobHandler; handler != nil {
+		return handler(jobID, reason)
+	}
+	if s.plugin == nil {
+		return nil, false, fmt.Errorf("plugin is not enabled")
+	}
+	return s.plugin.ExpireJob(jobID, reason)
+}
+
 // ListPluginActivities returns plugin job activities for monitoring.
 func (s *AdminServer) ListPluginActivities(jobType string, limit int) []adminplugin.JobActivity {
 	if s.plugin == nil {

weed/admin/dash/plugin_api.go

@@ -5,6 +5,7 @@ import (
 	"crypto/rand"
 	"encoding/hex"
 	"encoding/json"
+	"errors"
 	"fmt"
 	"io"
 	"net/http"
@@ -130,6 +131,47 @@ func (s *AdminServer) GetPluginJobDetailAPI(w http.ResponseWriter, r *http.Reque
 	writeJSON(w, http.StatusOK, detail)
 }
 
+// ExpirePluginJobAPI marks a job as failed so it no longer blocks scheduling.
+func (s *AdminServer) ExpirePluginJobAPI(w http.ResponseWriter, r *http.Request) {
+	jobID := strings.TrimSpace(mux.Vars(r)["jobId"])
+	if jobID == "" {
+		writeJSONError(w, http.StatusBadRequest, "jobId is required")
+		return
+	}
+
+	var req struct {
+		Reason string `json:"reason"`
+	}
+
+	if err := decodeJSONBody(newJSONMaxReader(w, r), &req); err != nil && err != io.EOF {
+		writeJSONError(w, http.StatusBadRequest, "invalid request body: "+err.Error())
+		return
+	}
+
+	job, expired, err := s.ExpirePluginJob(jobID, req.Reason)
+	if err != nil {
+		if errors.Is(err, plugin.ErrJobNotFound) {
+			writeJSONError(w, http.StatusNotFound, err.Error())
+			return
+		}
+		writeJSONError(w, http.StatusInternalServerError, err.Error())
+		return
+	}
+
+	response := map[string]interface{}{
+		"job_id":  jobID,
+		"expired": expired,
+	}
+	if job != nil {
+		response["job"] = job
+	}
+	if !expired {
+		response["message"] = "job is not active"
+	}
+
+	writeJSON(w, http.StatusOK, response)
+}
+
 // GetPluginActivitiesAPI returns recent plugin activities.
 func (s *AdminServer) GetPluginActivitiesAPI(w http.ResponseWriter, r *http.Request) {
 	query := r.URL.Query()

weed/admin/dash/plugin_api_test.go

@@ -1,11 +1,120 @@
 package dash
 
 import (
+	"encoding/json"
+	"errors"
+	"io"
+	"net/http"
+	"net/http/httptest"
+	"strings"
 	"testing"
 
+	"github.com/gorilla/mux"
+	"github.com/seaweedfs/seaweedfs/weed/admin/plugin"
 	"github.com/seaweedfs/seaweedfs/weed/pb/plugin_pb"
 )
 
+func TestExpirePluginJobAPI(t *testing.T) {
+	makeRequest := func(adminServer *AdminServer, jobID string, body io.Reader) *httptest.ResponseRecorder {
+		req := httptest.NewRequest(http.MethodPost, "/api/plugin/jobs/"+jobID+"/expire", body)
+		req = mux.SetURLVars(req, map[string]string{"jobId": jobID})
+		recorder := httptest.NewRecorder()
+		adminServer.ExpirePluginJobAPI(recorder, req)
+		return recorder
+	}
+
+	t.Run("empty job id", func(t *testing.T) {
+		recorder := makeRequest(&AdminServer{}, "", nil)
+		if recorder.Code != http.StatusBadRequest {
+			t.Fatalf("expected 400, got %d", recorder.Code)
+		}
+	})
+
+	t.Run("invalid json", func(t *testing.T) {
+		recorder := makeRequest(&AdminServer{}, "job-id", strings.NewReader("{"))
+		if recorder.Code != http.StatusBadRequest {
+			t.Fatalf("expected 400, got %d", recorder.Code)
+		}
+	})
+
+	t.Run("job not found", func(t *testing.T) {
+		adminServer := &AdminServer{
+			expireJobHandler: func(jobID, reason string) (*plugin.TrackedJob, bool, error) {
+				return nil, false, plugin.ErrJobNotFound
+			},
+		}
+		recorder := makeRequest(adminServer, "missing", strings.NewReader(`{"reason":"nope"}`))
+		if recorder.Code != http.StatusNotFound {
+			t.Fatalf("expected 404, got %d", recorder.Code)
+		}
+		var payload map[string]any
+		if err := json.Unmarshal(recorder.Body.Bytes(), &payload); err != nil {
+			t.Fatalf("failed to unmarshal body: %v", err)
+		}
+		if payload["error"] == nil {
+			t.Fatalf("expected error payload, got %v", payload)
+		}
+	})
+
+	t.Run("successful expire", func(t *testing.T) {
+		expected := &plugin.TrackedJob{JobID: "foo", State: "assigned"}
+		adminServer := &AdminServer{
+			expireJobHandler: func(jobID, reason string) (*plugin.TrackedJob, bool, error) {
+				if jobID != "foo" {
+					return nil, false, errors.New("unexpected")
+				}
+				return expected, true, nil
+			},
+		}
+		recorder := makeRequest(adminServer, "foo", strings.NewReader(`{"reason":"cleanup"}`))
+		if recorder.Code != http.StatusOK {
+			t.Fatalf("expected 200, got %d", recorder.Code)
+		}
+		var payload map[string]any
+		if err := json.Unmarshal(recorder.Body.Bytes(), &payload); err != nil {
+			t.Fatalf("failed to decode payload: %v", err)
+		}
+		if payload["job_id"] != "foo" {
+			t.Fatalf("expected job_id foo, got %v", payload["job_id"])
+		}
+		if expired, ok := payload["expired"].(bool); !ok || !expired {
+			t.Fatalf("expected expired=true, got %v", payload["expired"])
+		}
+		jobData, ok := payload["job"].(map[string]any)
+		if !ok || jobData["job_id"] != "foo" {
+			t.Fatalf("expected job info with job_id, got %v", payload["job"])
+		}
+	})
+
+	t.Run("non-active job", func(t *testing.T) {
+		adminServer := &AdminServer{
+			expireJobHandler: func(jobID, reason string) (*plugin.TrackedJob, bool, error) {
+				return nil, false, nil
+			},
+		}
+		recorder := makeRequest(adminServer, "bar", strings.NewReader(`{"reason":"ignore"}`))
+		if recorder.Code != http.StatusOK {
+			t.Fatalf("expected 200, got %d", recorder.Code)
+		}
+		var payload map[string]any
+		if err := json.Unmarshal(recorder.Body.Bytes(), &payload); err != nil {
+			t.Fatalf("failed to decode payload: %v", err)
+		}
+		if payload["job_id"] != "bar" {
+			t.Fatalf("expected job_id bar, got %v", payload["job_id"])
+		}
+		if expired, ok := payload["expired"].(bool); !ok || expired {
+			t.Fatalf("expected expired=false, got %v", payload["expired"])
+		}
+		if payload["message"] != "job is not active" {
+			t.Fatalf("expected message job is not active, got %v", payload["message"])
+		}
+		if _, exists := payload["job"]; exists {
+			t.Fatalf("expected no job payload for non-active job, got %v", payload["job"])
+		}
+	})
+}
+
 func TestBuildJobSpecFromProposalDoesNotReuseProposalID(t *testing.T) {
 	t.Parallel()
 

weed/admin/handlers/admin_handlers.go

@@ -242,6 +242,7 @@ func (h *AdminHandlers) registerAPIRoutes(api *mux.Router, enforceWrite bool) {
 	pluginApi.Handle("/job-types/{jobType}/detect", wrapWrite(h.adminServer.TriggerPluginDetectionAPI)).Methods(http.MethodPost)
 	pluginApi.Handle("/job-types/{jobType}/run", wrapWrite(h.adminServer.RunPluginJobTypeAPI)).Methods(http.MethodPost)
 	pluginApi.Handle("/jobs/execute", wrapWrite(h.adminServer.ExecutePluginJobAPI)).Methods(http.MethodPost)
+	pluginApi.Handle("/jobs/{jobId}/expire", wrapWrite(h.adminServer.ExpirePluginJobAPI)).Methods(http.MethodPost)
 
 	mqApi := api.PathPrefix("/mq").Subrouter()
 	mqApi.HandleFunc("/topics/{namespace}/{topic}", h.mqHandlers.GetTopicDetailsAPI).Methods(http.MethodGet)

weed/admin/handlers/admin_handlers_routes_test.go

@@ -21,6 +21,9 @@ func TestSetupRoutes_RegistersPluginSchedulerStatesAPI_NoAuth(t *testing.T) {
 	if !hasRoute(router, http.MethodGet, "/api/plugin/jobs/example/detail") {
 		t.Fatalf("expected GET /api/plugin/jobs/:jobId/detail to be registered in no-auth mode")
 	}
+	if !hasRoute(router, http.MethodPost, "/api/plugin/jobs/example/expire") {
+		t.Fatalf("expected POST /api/plugin/jobs/:jobId/expire to be registered in no-auth mode")
+	}
 }
 
 func TestSetupRoutes_RegistersPluginSchedulerStatesAPI_WithAuth(t *testing.T) {
@@ -34,6 +37,9 @@ func TestSetupRoutes_RegistersPluginSchedulerStatesAPI_WithAuth(t *testing.T) {
 	if !hasRoute(router, http.MethodGet, "/api/plugin/jobs/example/detail") {
 		t.Fatalf("expected GET /api/plugin/jobs/:jobId/detail to be registered in auth mode")
 	}
+	if !hasRoute(router, http.MethodPost, "/api/plugin/jobs/example/expire") {
+		t.Fatalf("expected POST /api/plugin/jobs/:jobId/expire to be registered in auth mode")
+	}
 }
 
 func TestSetupRoutes_RegistersPluginPages_NoAuth(t *testing.T) {

weed/admin/plugin/plugin.go

@@ -80,6 +80,8 @@ type Plugin struct {
 
 	jobsMu sync.RWMutex
 	jobs   map[string]*TrackedJob
+	// serialize stale job cleanup to avoid duplicate expirations
+	staleJobsMu sync.Mutex
 
 	jobDetailsMu sync.Mutex
 

weed/admin/plugin/plugin_monitor.go

@@ -2,6 +2,7 @@ package plugin
 
 import (
 	"encoding/json"
+	"fmt"
 	"sort"
 	"strings"
 	"time"
@@ -9,12 +10,18 @@ import (
 	"github.com/seaweedfs/seaweedfs/weed/glog"
 	"github.com/seaweedfs/seaweedfs/weed/pb/plugin_pb"
 	"google.golang.org/protobuf/encoding/protojson"
+	"google.golang.org/protobuf/types/known/timestamppb"
 )
 
 const (
 	maxTrackedJobsTotal = 1000
 	maxActivityRecords  = 4000
 	maxRelatedJobs      = 100
+
+	// stale active jobs block dedupe and scheduling; use generous defaults to
+	// avoid expiring legitimate long-running tasks.
+	defaultStaleActiveJobTimeout    = 24 * time.Hour
+	defaultOrphanedActiveJobTimeout = 15 * time.Minute
 )
 
 var (
@@ -23,6 +30,14 @@ var (
 	StateCanceled  = strings.ToLower(plugin_pb.JobState_JOB_STATE_CANCELED.String())
 )
 
+type activeJobSnapshot struct {
+	jobID      string
+	jobType    string
+	workerID   string
+	requestID  string
+	lastUpdate time.Time
+}
+
 // activityLess reports whether activity a occurred after activity b (newest-first order).
 // A nil OccurredAt is treated as the zero time.
 func activityLess(a, b JobActivity) bool {
@@ -54,6 +69,13 @@ func (r *Plugin) loadPersistedMonitorState() error {
 			if strings.TrimSpace(job.JobID) == "" {
 				continue
 			}
+			if isActiveTrackedJobState(job.State) {
+				if detail, detailErr := r.store.LoadJobDetail(job.JobID); detailErr != nil {
+					glog.Warningf("Plugin failed to load detail snapshot for job %s: %v", job.JobID, detailErr)
+				} else if detail != nil {
+					mergeTerminalDetailIntoTracked(&job, detail)
+				}
+			}
 			// Backward compatibility: migrate older inline detail payloads
 			// out of tracked_jobs.json into dedicated per-job detail files.
 			if hasTrackedJobRichDetails(job) {
@@ -81,6 +103,265 @@ func (r *Plugin) loadPersistedMonitorState() error {
 	return nil
 }
 
+// ExpireJob marks an active job as failed so it no longer blocks scheduling.
+func (r *Plugin) ExpireJob(jobID, reason string) (*TrackedJob, bool, error) {
+	normalizedJobID := strings.TrimSpace(jobID)
+	if normalizedJobID == "" {
+		return nil, false, ErrJobNotFound
+	}
+
+	reason = strings.TrimSpace(reason)
+	if reason == "" {
+		reason = "job expired by admin request"
+	}
+
+	var jobType string
+	var requestID string
+	active := false
+
+	r.jobsMu.RLock()
+	if tracked := r.jobs[normalizedJobID]; tracked != nil {
+		jobType = tracked.JobType
+		requestID = tracked.RequestID
+		active = isActiveTrackedJobState(tracked.State)
+	}
+	r.jobsMu.RUnlock()
+
+	if jobType == "" || requestID == "" || !active {
+		if detail, err := r.store.LoadJobDetail(normalizedJobID); err != nil {
+			return nil, false, err
+		} else if detail != nil {
+			if jobType == "" {
+				jobType = detail.JobType
+			}
+			if requestID == "" {
+				requestID = detail.RequestID
+			}
+			if !active && isActiveTrackedJobState(detail.State) {
+				active = true
+			}
+		}
+	}
+
+	if jobType == "" {
+		return nil, false, ErrJobNotFound
+	}
+
+	if !active {
+		current, _ := r.GetTrackedJob(normalizedJobID)
+		if current == nil {
+			if detail, err := r.store.LoadJobDetail(normalizedJobID); err == nil && detail != nil {
+				clone := cloneTrackedJob(*detail)
+				current = &clone
+			}
+		}
+		return current, false, nil
+	}
+
+	now := time.Now().UTC()
+	r.handleJobCompleted(&plugin_pb.JobCompleted{
+		JobId:        normalizedJobID,
+		JobType:      jobType,
+		RequestId:    requestID,
+		Success:      false,
+		ErrorMessage: reason,
+		CompletedAt:  timestamppb.New(now),
+	})
+	r.appendActivity(JobActivity{
+		JobID:      normalizedJobID,
+		JobType:    jobType,
+		RequestID:  requestID,
+		Source:     "admin_expire",
+		Message:    reason,
+		Stage:      "expired",
+		OccurredAt: timeToPtr(now),
+	})
+
+	updated, _ := r.GetTrackedJob(normalizedJobID)
+	return updated, true, nil
+}
+
+// expireStaleJobs marks stale active jobs as failed so they stop blocking new work.
+func (r *Plugin) expireStaleJobs(now time.Time) int {
+	if now.IsZero() {
+		now = time.Now().UTC()
+	}
+
+	r.staleJobsMu.Lock()
+	defer r.staleJobsMu.Unlock()
+
+	snapshots := r.snapshotActiveJobs()
+	if len(snapshots) == 0 {
+		return 0
+	}
+
+	expired := 0
+	for _, snap := range snapshots {
+		if snap.lastUpdate.IsZero() {
+			continue
+		}
+		if stale, _, _ := r.evaluateStaleJob(now, snap.workerID, snap.lastUpdate); !stale {
+			continue
+		}
+
+		reason := r.confirmStaleReason(now, snap.jobID)
+		if reason == "" {
+			continue
+		}
+
+		r.handleJobCompleted(&plugin_pb.JobCompleted{
+			JobId:        snap.jobID,
+			JobType:      snap.jobType,
+			RequestId:    snap.requestID,
+			Success:      false,
+			ErrorMessage: reason,
+			CompletedAt:  timestamppb.New(now),
+		})
+		expired++
+	}
+
+	return expired
+}
+
+func (r *Plugin) snapshotActiveJobs() []activeJobSnapshot {
+	r.jobsMu.RLock()
+	defer r.jobsMu.RUnlock()
+
+	if len(r.jobs) == 0 {
+		return nil
+	}
+
+	out := make([]activeJobSnapshot, 0, len(r.jobs))
+	for _, job := range r.jobs {
+		if job == nil {
+			continue
+		}
+		if !isActiveTrackedJobState(job.State) {
+			continue
+		}
+		out = append(out, activeJobSnapshot{
+			jobID:      job.JobID,
+			jobType:    job.JobType,
+			workerID:   job.WorkerID,
+			requestID:  job.RequestID,
+			lastUpdate: jobLastUpdated(job),
+		})
+	}
+	return out
+}
+
+func jobLastUpdated(job *TrackedJob) time.Time {
+	if job == nil {
+		return time.Time{}
+	}
+	if job.UpdatedAt != nil && !job.UpdatedAt.IsZero() {
+		return *job.UpdatedAt
+	}
+	if job.CreatedAt != nil && !job.CreatedAt.IsZero() {
+		return *job.CreatedAt
+	}
+	return time.Time{}
+}
+
+func (r *Plugin) evaluateStaleJob(now time.Time, workerID string, lastUpdate time.Time) (bool, time.Duration, string) {
+	if lastUpdate.IsZero() {
+		return false, 0, ""
+	}
+
+	timeout := defaultStaleActiveJobTimeout
+	reason := fmt.Sprintf("job expired after %s without progress", timeout)
+
+	workerID = strings.TrimSpace(workerID)
+	if workerID == "" {
+		reason = fmt.Sprintf("job expired after %s without executor assignment", timeout)
+	} else if !r.isWorkerAvailable(workerID) {
+		timeout = defaultOrphanedActiveJobTimeout
+		reason = fmt.Sprintf("job expired after %s without worker heartbeat (worker=%s)", timeout, workerID)
+	}
+
+	if now.Sub(lastUpdate) < timeout {
+		return false, timeout, reason
+	}
+	return true, timeout, reason
+}
+
+func (r *Plugin) confirmStaleReason(now time.Time, jobID string) string {
+	r.jobsMu.RLock()
+	job := r.jobs[jobID]
+	if job == nil || !isActiveTrackedJobState(job.State) {
+		r.jobsMu.RUnlock()
+		return ""
+	}
+	lastUpdate := jobLastUpdated(job)
+	workerID := job.WorkerID
+	r.jobsMu.RUnlock()
+
+	stale, _, reason := r.evaluateStaleJob(now, workerID, lastUpdate)
+	if !stale {
+		return ""
+	}
+	return reason
+}
+
+func (r *Plugin) isWorkerAvailable(workerID string) bool {
+	workerID = strings.TrimSpace(workerID)
+	if workerID == "" {
+		return false
+	}
+	_, ok := r.registry.Get(workerID)
+	return ok
+}
+
+func isTerminalTrackedJobState(state string) bool {
+	normalized := strings.ToLower(strings.TrimSpace(state))
+	switch normalized {
+	case StateSucceeded, StateFailed, StateCanceled:
+		return true
+	default:
+		return false
+	}
+}
+
+func mergeTerminalDetailIntoTracked(tracked *TrackedJob, detail *TrackedJob) {
+	if tracked == nil || detail == nil {
+		return
+	}
+	if !isTerminalTrackedJobState(detail.State) {
+		return
+	}
+	if !isActiveTrackedJobState(tracked.State) {
+		return
+	}
+
+	if detail.State != "" {
+		tracked.State = detail.State
+	}
+	if detail.Progress != 0 {
+		tracked.Progress = detail.Progress
+	}
+	if detail.Stage != "" {
+		tracked.Stage = detail.Stage
+	}
+	if detail.Message != "" {
+		tracked.Message = detail.Message
+	}
+	if detail.ErrorMessage != "" {
+		tracked.ErrorMessage = detail.ErrorMessage
+	}
+	if detail.ResultSummary != "" {
+		tracked.ResultSummary = detail.ResultSummary
+	}
+	if detail.CompletedAt != nil && !detail.CompletedAt.IsZero() {
+		tracked.CompletedAt = detail.CompletedAt
+	}
+	if detail.UpdatedAt != nil && !detail.UpdatedAt.IsZero() {
+		tracked.UpdatedAt = detail.UpdatedAt
+	}
+	if tracked.UpdatedAt == nil && tracked.CompletedAt != nil {
+		tracked.UpdatedAt = tracked.CompletedAt
+	}
+}
+
 func (r *Plugin) ListTrackedJobs(jobType string, state string, limit int) []TrackedJob {
 	r.jobsMu.RLock()
 	defer r.jobsMu.RUnlock()

weed/admin/plugin/plugin_scheduler.go

@@ -61,6 +61,8 @@ func (r *Plugin) schedulerLoop() {
 }
 
 func (r *Plugin) runSchedulerTick() {
+	r.expireStaleJobs(time.Now().UTC())
+
 	jobTypes := r.registry.DetectableJobTypes()
 	if len(jobTypes) == 0 {
 		return
@@ -839,11 +841,16 @@ func waitForShutdownOrTimer(shutdown <-chan struct{}, duration time.Duration) bo
 	}
 }
 
+// filterProposalsWithActiveJobs removes proposals whose dedupe keys already have active jobs.
+// It first expires stale tracked jobs via expireStaleJobs, which can mutate scheduler state,
+// so callers should treat this method as a stateful operation.
 func (r *Plugin) filterProposalsWithActiveJobs(jobType string, proposals []*plugin_pb.JobProposal) ([]*plugin_pb.JobProposal, int) {
 	if len(proposals) == 0 {
 		return proposals, 0
 	}
 
+	r.expireStaleJobs(time.Now().UTC())
+
 	activeKeys := make(map[string]struct{})
 	r.jobsMu.RLock()
 	for _, job := range r.jobs {

weed/admin/plugin/types.go

@@ -1,6 +1,9 @@
 package plugin
 
-import "time"
+import (
+	"errors"
+	"time"
+)
 
 const (
 	// Keep exactly the last 10 successful and last 10 error runs per job type.
@@ -8,6 +11,8 @@ const (
 	MaxErrorRunHistory      = 10
 )
 
+var ErrJobNotFound = errors.New("job not found")
+
 type RunOutcome string
 
 const (

weed/admin/view/app/plugin.templ

@@ -485,6 +485,11 @@ templ Plugin(page string) {
             <div class="modal-content">
                 <div class="modal-header">
                     <h5 class="modal-title" id="plugin-job-detail-modal-label"><i class="fas fa-file-alt me-2"></i>Job Detail</h5>
+                    <div class="ms-auto me-2">
+                        <button type="button" class="btn btn-outline-danger btn-sm" id="plugin-expire-job-btn" disabled>
+                            <i class="fas fa-stop-circle me-1"></i>Expire Job
+                        </button>
+                    </div>
                     <button type="button" class="btn-close" data-bs-dismiss="modal" aria-label="Close"></button>
                 </div>
                 <div class="modal-body" id="plugin-job-detail-content">
@@ -1073,6 +1078,66 @@ templ Plugin(page string) {
                 return html;
             }
 
+            function isActiveJobState(candidateState) {
+                var jobState = candidateState;
+                if (candidateState && typeof candidateState === 'object' && candidateState.state !== undefined) {
+                    jobState = candidateState.state;
+                }
+                var st = String(jobState || '').toLowerCase();
+                return st === 'job_state_pending' || st === 'job_state_assigned' || st === 'job_state_running' ||
+                    st === 'pending' || st === 'assigned' || st === 'running' || st === 'in_progress';
+            }
+
+            function setExpireButtonState(job) {
+                var expireBtn = document.getElementById('plugin-expire-job-btn');
+                if (!expireBtn) {
+                    return;
+                }
+                var jobID = job && job.job_id ? String(job.job_id) : '';
+                var active = isActiveJobState(job);
+                expireBtn.setAttribute('data-job-id', jobID);
+                expireBtn.disabled = !jobID || !active;
+                if (!jobID) {
+                    expireBtn.title = 'Select a job to expire.';
+                } else if (!active) {
+                    expireBtn.title = 'Job is not active.';
+                } else {
+                    expireBtn.title = 'Expire job to unblock scheduling.';
+                }
+            }
+
+            async function expireJob(jobID) {
+                var normalizedJobID = String(jobID || '').trim();
+                if (!normalizedJobID) {
+                    return;
+                }
+
+                var reason = window.prompt('Expire job ' + normalizedJobID + '? Optional reason:', 'job expired by admin request');
+                if (reason === null) {
+                    return;
+                }
+
+                var expireBtn = document.getElementById('plugin-expire-job-btn');
+                if (expireBtn) {
+                    expireBtn.disabled = true;
+                }
+
+                try {
+                    var response = await pluginRequest('POST', '/api/plugin/jobs/' + encodePath(normalizedJobID) + '/expire', {
+                        reason: reason,
+                    });
+                    if (response && response.expired === false) {
+                        notify(response.message || 'Job is not active.', 'info');
+                    } else {
+                        notify('Job expired: ' + normalizedJobID, 'success');
+                    }
+                    await refreshJobsAndActivities();
+                    await openJobDetail(normalizedJobID);
+                } catch (e) {
+                    notify('Failed to expire job: ' + e.message, 'error');
+                }
+            }
+
             async function openJobDetail(jobID) {
                 var normalizedJobID = String(jobID || '').trim();
                 if (!normalizedJobID) {
@@ -1093,10 +1158,12 @@ templ Plugin(page string) {
                     modal.show();
                 }
 
+                setExpireButtonState(null);
                 contentRoot.innerHTML = '<div class="text-muted">Loading job detail...</div>';
                 try {
                     var detail = await pluginRequest('GET', '/api/plugin/jobs/' + encodePath(normalizedJobID) + '/detail?activity_limit=500&related_limit=20');
                     var job = (detail && detail.job) ? detail.job : {};
+                    setExpireButtonState(job);
                     var runRecord = detail && detail.run_record ? detail.run_record : null;
                     var activities = (detail && Array.isArray(detail.activities)) ? detail.activities : [];
                     var relatedJobs = (detail && Array.isArray(detail.related_jobs)) ? detail.related_jobs : [];
@@ -1197,6 +1264,7 @@ templ Plugin(page string) {
 
                     contentRoot.innerHTML = html;
                 } catch (e) {
+                    setExpireButtonState(null);
                     contentRoot.innerHTML = '<div class="alert alert-danger mb-0">Failed to load job detail: ' + escapeHtml(e.message) + '</div>';
                 }
             }
@@ -1238,8 +1306,7 @@ templ Plugin(page string) {
                 var allActivities = Array.isArray(state.allActivities) ? state.allActivities : [];
 
                 var activeCount = allJobs.filter(function(job) {
-                    var st = String(job.state || '').toLowerCase();
+                    return isActiveJobState(job);
-                    return st === 'job_state_pending' || st === 'job_state_assigned' || st === 'job_state_running' || st === 'pending' || st === 'assigned' || st === 'running' || st === 'in_progress';
                 }).length;
 
                 document.getElementById('plugin-status-workers').textContent = String(state.workers.length);
@@ -1265,8 +1332,7 @@ templ Plugin(page string) {
                     if (!jobType) {
                         continue;
                     }
-                    var st = String(job.state || '').toLowerCase();
+                    var isActive = isActiveJobState(job);
-                    var isActive = st === 'job_state_pending' || st === 'job_state_assigned' || st === 'job_state_running' || st === 'pending' || st === 'assigned' || st === 'running' || st === 'in_progress';
                     if (!isActive) {
                         continue;
                     }
@@ -2778,6 +2844,17 @@ templ Plugin(page string) {
                     });
                 }
 
+                var expireBtn = document.getElementById('plugin-expire-job-btn');
+                if (expireBtn) {
+                    expireBtn.addEventListener('click', function() {
+                        var jobID = String(expireBtn.getAttribute('data-job-id') || '').trim();
+                        if (!jobID) {
+                            return;
+                        }
+                        expireJob(jobID);
+                    });
+                }
+
                 document.getElementById('plugin-refresh-all-btn').addEventListener('click', function() {
                     refreshAll();
                 });