fix(monitor): 修复取件死信无法重排与列表 NULL 标志导致的 500

- platform/postgres: listTestPublicationsSQL 的 can_start/can_retry 在 LEFT JOIN
  内容对象为空时算出 NULL,Go 侧 bool 扫描失败使整个列表接口 500,前端三个 lane
  全空(表现为“待入库 0 / 全部任务 0”)。改为 COALESCE(..., false)。
- 新增 POST /api/v1/job-failures/{id}/requeue:取件任务的幂等键绑在版本上,
  失败那次已占住它,重新标注同一文件不会再入队;现在按新轮次 (…:r<n>) 重排、
  复位意图取件状态并把死信标记为 requeued(幂等信封 + 审计字段)。
- web: 死信区对 artifact.fetch 显示「重新入队」;系统设置新增「知识库绑定」分区,
  挂上一直没被渲染的 RoutingBindingsSection,导入目标库可在页面上切换。
- 测试:真实 PostgreSQL 集成测试(NULL 折叠 + 重排轮次/幂等/拒绝)、前端死信按钮
  与设置分区渲染测试。
This commit is contained in:
hermes
2026-10-08 12:09:34 +08:00
parent 22a6bc5004
commit cddc4c90d1
12 changed files with 495 additions and 8 deletions
+50
View File
@@ -96,6 +96,7 @@ type Store interface {
ListJobFailures(ctx context.Context, filter governance.JobFailureFilter) (governance.JobFailurePage, error)
RetryPublicationsBatchWithOperation(ctx context.Context, request governance.RetryPublicationsBatchRequest, operationID string) (governance.RetryPublicationsBatchResult, bool, error)
CancelPublicationsBatchWithOperation(ctx context.Context, request governance.CancelPublicationsBatchRequest, operationID string) (governance.RetryPublicationsBatchResult, bool, error)
RequeueArtifactFetchWithOperation(ctx context.Context, request governance.RequeueArtifactFetchRequest, operationID string) (governance.RequeueArtifactFetchResult, bool, error)
}
type Config struct {
@@ -394,6 +395,9 @@ func NewHandler(config Config) (http.Handler, error) {
mux.HandleFunc("GET /api/v1/job-failures", func(w http.ResponseWriter, r *http.Request) {
listJobFailures(w, r, config.Store)
})
mux.HandleFunc("POST /api/v1/job-failures/{id}/requeue", func(w http.ResponseWriter, r *http.Request) {
requeueJobFailure(w, r, config)
})
if config.Auth != nil {
if config.Auth.Store == nil {
return nil, errors.New("auth configuration requires a store")
@@ -2016,6 +2020,52 @@ func listJobFailures(w http.ResponseWriter, r *http.Request, store Store) {
writeJSON(w, http.StatusOK, map[string]any{"data": page})
}
// requeueJobFailure re-enqueues one terminal 源头取件 (artifact.fetch) dead
// letter. 上传解析 dead letters keep their publication-keyed retry path; a
// fetch failure carries no publication id, so without this route the file is
// stuck forever after a single terminal fetch failure (expired source session).
func requeueJobFailure(w http.ResponseWriter, r *http.Request, config Config) {
operationID, ok := requireOperationID(w, r)
if !ok {
return
}
failureID := strings.TrimSpace(r.PathValue("id"))
if failureID == "" || len(failureID) > 64 || !validCatalogID(failureID) {
writeJSON(w, http.StatusBadRequest, map[string]string{"error": "invalid failure id"})
return
}
var input struct {
Reason string `json:"reason"`
}
if err := decodeStrictJSON(r, &input); err != nil {
writeJSON(w, http.StatusBadRequest, map[string]string{"error": "invalid requeue request"})
return
}
reason := strings.TrimSpace(input.Reason)
if reason == "" || len(reason) > 500 {
writeJSON(w, http.StatusBadRequest, map[string]string{"error": "reason is required"})
return
}
result, replayed, err := config.Store.RequeueArtifactFetchWithOperation(r.Context(), governance.RequeueArtifactFetchRequest{
FailureID: failureID,
OperatorRef: actorFor(r),
Reason: reason,
Now: config.Now().UTC(),
}, operationID)
switch {
case errors.Is(err, postgresstore.ErrJobFailureNotRequeueable):
writeJSON(w, http.StatusConflict, map[string]string{"error": "该失败记录不是可重排的源头取件任务,或已经处理过"})
return
case errors.Is(err, postgresstore.ErrOperationRequestConflict):
writeJSON(w, http.StatusConflict, map[string]string{"error": "Idempotency-Key conflicts with a different request"})
return
case err != nil:
writeJSON(w, http.StatusInternalServerError, map[string]string{"error": "could not requeue artifact fetch"})
return
}
writeJSON(w, http.StatusOK, map[string]any{"data": result, "replayed": replayed})
}
func retryPublicationsBatch(w http.ResponseWriter, r *http.Request, config Config) {
operationID, ok := requireOperationID(w, r)
if !ok {
+11
View File
@@ -1515,6 +1515,11 @@ type fakeJobStore struct {
cancelBatchErr error
retryBatchReplayed bool
retryBatchErr error
requeueFailureRequest *governance.RequeueArtifactFetchRequest
requeueFailureOperation string
requeueFailureResult governance.RequeueArtifactFetchResult
requeueFailureReplayed bool
requeueFailureErr error
enqueued jobs.EnqueueRequest
enqueueResult *jobs.Job
listResult []jobs.Job
@@ -1925,6 +1930,12 @@ func (s *fakeJobStore) CancelPublicationsBatchWithOperation(_ context.Context, r
return s.cancelBatchResult, s.cancelBatchReplayed, s.cancelBatchErr
}
func (s *fakeJobStore) RequeueArtifactFetchWithOperation(_ context.Context, request governance.RequeueArtifactFetchRequest, operationID string) (governance.RequeueArtifactFetchResult, bool, error) {
s.requeueFailureRequest = &request
s.requeueFailureOperation = operationID
return s.requeueFailureResult, s.requeueFailureReplayed, s.requeueFailureErr
}
func (s *fakeJobStore) ListScopeRules(_ context.Context, connectionID string) ([]governance.ScopeRule, error) {
s.scopeRuleConnectionID = connectionID
return s.scopeRules, s.scopeRuleErr