Configurable workers and move sweeper job to job infra (#35007)

e2e-tests/playwright/lib/src/server/default_config.ts

@@ -849,6 +849,7 @@ const defaultServerConfig: AdminConfig = {
             APIKey: '',
         },
         TargetLanguages: ['en'],
+        Workers: 4,
         TimeoutMs: 5000,
         Agents: {
             LLMServiceID: '',

server/channels/app/server.go

@@ -1488,6 +1488,12 @@ func (s *Server) initJobs() {
 		s.Jobs.RegisterJobType(model.JobTypePushProxyAuth, builder.MakeWorker(), builder.MakeScheduler())
 	}
 
+	if s.AutoTranslation != nil {
+		s.Jobs.RegisterJobType(model.JobTypeAutoTranslationRecovery,
+			s.AutoTranslation.MakeWorker(),
+			s.AutoTranslation.MakeScheduler())
+	}
+
 	s.Jobs.RegisterJobType(
 		model.JobTypeMigrations,
 		migrations.MakeWorker(s.Jobs, s.Store()),

server/einterfaces/autotranslation.go

@@ -7,6 +7,7 @@ import (
 	"context"
 
 	"github.com/mattermost/mattermost/server/public/model"
+	ejobs "github.com/mattermost/mattermost/server/v8/einterfaces/jobs"
 )
 
 // AutoTranslationInterface defines the enterprise advanced auto-translation functionality.
@@ -75,4 +76,12 @@ type AutoTranslationInterface interface {
 
 	// Shutdown gracefully shuts down the auto-translation service.
 	Shutdown() error
+
+	// MakeWorker creates a worker for the autotranslation recovery sweep job.
+	// The worker picks up stuck translations and re-queues them for processing.
+	MakeWorker() model.Worker
+
+	// MakeScheduler creates a scheduler for the autotranslation recovery sweep job.
+	// The scheduler runs periodically to detect stuck translations.
+	MakeScheduler() ejobs.Scheduler
 }

server/einterfaces/mocks/AutoTranslationInterface.go

@@ -7,6 +7,8 @@ package mocks
 import (
 	context "context"
 
+	jobs "github.com/mattermost/mattermost/server/v8/einterfaces/jobs"
+
 	mock "github.com/stretchr/testify/mock"
 
 	model "github.com/mattermost/mattermost/server/public/model"
@@ -214,6 +216,46 @@ func (_m *AutoTranslationInterface) IsUserEnabled(channelID string, userID strin
 	return r0, r1
 }
 
+// MakeScheduler provides a mock function with no fields
+func (_m *AutoTranslationInterface) MakeScheduler() jobs.Scheduler {
+	ret := _m.Called()
+
+	if len(ret) == 0 {
+		panic("no return value specified for MakeScheduler")
+	}
+
+	var r0 jobs.Scheduler
+	if rf, ok := ret.Get(0).(func() jobs.Scheduler); ok {
+		r0 = rf()
+	} else {
+		if ret.Get(0) != nil {
+			r0 = ret.Get(0).(jobs.Scheduler)
+		}
+	}
+
+	return r0
+}
+
+// MakeWorker provides a mock function with no fields
+func (_m *AutoTranslationInterface) MakeWorker() model.Worker {
+	ret := _m.Called()
+
+	if len(ret) == 0 {
+		panic("no return value specified for MakeWorker")
+	}
+
+	var r0 model.Worker
+	if rf, ok := ret.Get(0).(func() model.Worker); ok {
+		r0 = rf()
+	} else {
+		if ret.Get(0) != nil {
+			r0 = ret.Get(0).(model.Worker)
+		}
+	}
+
+	return r0
+}
+
 // SetChannelEnabled provides a mock function with given fields: channelID, enabled
 func (_m *AutoTranslationInterface) SetChannelEnabled(channelID string, enabled bool) *model.AppError {
 	ret := _m.Called(channelID, enabled)

server/i18n/en.json

@@ -9904,6 +9904,10 @@
     "id": "model.config.is_valid.autotranslation.timeout.app_error",
     "translation": "Invalid timeout for autotranslation settings. Must be a positive number."
   },
+  {
+    "id": "model.config.is_valid.autotranslation.workers.app_error",
+    "translation": "Workers must be between 1 and 32."
+  },
   {
     "id": "model.config.is_valid.cache_type.app_error",
     "translation": "Cache type must be either lru or redis."

server/public/model/config.go

@@ -2787,6 +2787,7 @@ type AutoTranslationSettings struct {
 	Enable          *bool                           `access:"site_localization,cloud_restrictable"`
 	Provider        *string                         `access:"site_localization,cloud_restrictable"`
 	TargetLanguages *[]string                       `access:"site_localization,cloud_restrictable"`
+	Workers         *int                            `access:"site_localization,cloud_restrictable"`
 	TimeoutMs       *int                            `access:"site_localization,cloud_restrictable"`
 	LibreTranslate  *LibreTranslateProviderSettings `access:"site_localization,cloud_restrictable"`
 	Agents          *AgentsProviderSettings         `access:"site_localization,cloud_restrictable"`
@@ -2815,6 +2816,10 @@ func (s *AutoTranslationSettings) SetDefaults() {
 		s.TargetLanguages = &[]string{"en"}
 	}
 
+	if s.Workers == nil {
+		s.Workers = NewPointer(4)
+	}
+
 	if s.TimeoutMs == nil {
 		s.TimeoutMs = NewPointer(5000)
 	}
@@ -4823,6 +4828,11 @@ func (s *AutoTranslationSettings) isValid() *AppError {
 		return NewAppError("Config.IsValid", "model.config.is_valid.autotranslation.timeout.app_error", nil, "", http.StatusBadRequest)
 	}
 
+	// Validate workers if set (must be between 1 and 32)
+	if s.Workers != nil && (*s.Workers < 1 || *s.Workers > 32) {
+		return NewAppError("Config.IsValid", "model.config.is_valid.autotranslation.workers.app_error", nil, "", http.StatusBadRequest)
+	}
+
 	return nil
 }
 

server/public/model/job.go

@@ -47,6 +47,7 @@ const (
 	JobTypePushProxyAuth                 = "push_proxy_auth"
 	JobTypeRecap                         = "recap"
 	JobTypeDeleteExpiredPosts            = "delete_expired_posts"
+	JobTypeAutoTranslationRecovery       = "autotranslation_recovery"
 
 	JobStatusPending         = "pending"
 	JobStatusInProgress      = "in_progress"

webapp/platform/types/src/config.ts

@@ -755,6 +755,7 @@ export type LocalizationSettings = {
 export type AutoTranslationSettings = {
     Enable: boolean;
     TargetLanguages: string[];
+    Workers: number;
     Provider: '' | 'libretranslate' | 'agents';
     LibreTranslate: {
         URL: string;