feat(server): hot-reload config store and reloadable infra clients

- Replace single Load() with a watching Store that re-reads config files
  (and config.local.yaml) and fans out a ReloadEvent with a per-field diff
  so consumers can decide whether the change is hot-applicable or requires
  a process restart.
- Wrap llm, retrieval, vector store, and object storage clients in
  Reloadable* shells so the bootstrap can swap their underlying impls when
  config changes without re-instantiating handlers.
- Make jwt.Manager and ops TokenIssuer mutable under a lock so secrets and
  TTLs can be rotated live; thread default plan code through a setter on
  the ops AdminUserService.
- Wire ConfigStore through bootstrap and every cmd/main.go, scheduler /
  worker / tenant-api / ops-api start the watcher; services and handlers
  take a config.Provider so they always read current values for things
  like generation.stream_enabled, scheduler dispatch, retrieval, etc.
- Switch shared/config decoding off viper to a Kratos-derived runtime
  package so env placeholders (\${VAR:default}) resolve consistently and
  the same source machinery powers both the loader and the watcher.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
2026-05-01 16:01:23 +08:00
parent ce2d8a2907
commit 618399f86d
61 changed files with 3186 additions and 496 deletions
@@ -34,6 +34,7 @@ type ArticleGenerationWorker struct {
logger *zap.Logger
retryInterval time.Duration
processTimeout time.Duration
configProvider config.Provider
consumerPrefix string
workerConcurrency int
}
@@ -53,11 +54,6 @@ func NewArticleGenerationWorker(
workerConcurrency = 1
}
processTimeout := cfg.ArticleTimeout + 2*time.Minute
if processTimeout <= 2*time.Minute {
processTimeout = 10 * time.Minute
}
return &ArticleGenerationWorker{
pool: pool,
rabbitMQ: rabbitMQClient,
@@ -67,12 +63,17 @@ func NewArticleGenerationWorker(
kolWorker: kolWorker,
logger: logger,
retryInterval: defaultArticleGenerationWorkerRetryInterval,
processTimeout: processTimeout,
processTimeout: articleGenerationProcessTimeout(cfg),
consumerPrefix: fmt.Sprintf("worker-generate-%d", os.Getpid()),
workerConcurrency: workerConcurrency,
}
}
func (w *ArticleGenerationWorker) WithConfigProvider(provider config.Provider) *ArticleGenerationWorker {
w.configProvider = provider
return w
}
func (w *ArticleGenerationWorker) Start(ctx context.Context) {
if w == nil || w.pool == nil || w.rabbitMQ == nil {
return
@@ -134,7 +135,7 @@ func (w *ArticleGenerationWorker) handleDelivery(parent context.Context, deliver
return
}
ctx, cancel := context.WithTimeout(parent, w.processTimeout)
ctx, cancel := context.WithTimeout(parent, w.runtimeProcessTimeout())
defer cancel()
task, terminal, err := w.loadTask(ctx, envelope.TaskID)
@@ -205,6 +206,26 @@ func (w *ArticleGenerationWorker) handleDelivery(parent context.Context, deliver
}
}
func (w *ArticleGenerationWorker) runtimeProcessTimeout() time.Duration {
if w != nil && w.configProvider != nil {
if cfg := w.configProvider.Current(); cfg != nil {
return articleGenerationProcessTimeout(cfg.Generation)
}
}
if w != nil && w.processTimeout > 0 {
return w.processTimeout
}
return articleGenerationProcessTimeout(config.GenerationConfig{})
}
func articleGenerationProcessTimeout(cfg config.GenerationConfig) time.Duration {
processTimeout := cfg.ArticleTimeout + 2*time.Minute
if processTimeout <= 2*time.Minute {
return 10 * time.Minute
}
return processTimeout
}
func (w *ArticleGenerationWorker) handleTaskFailure(ctx context.Context, task tenantapp.ClaimedGenerationTask, stage string, taskErr error) {
task.TaskType = strings.TrimSpace(task.TaskType)
switch {