{"record":{"id":"17271810b64b2c23","repo":"bytedance/deer-flow","slug":"scheduler-multi-instance-true-requires-run-events","errorCode":null,"errorMessage":"scheduler.multi_instance=true requires run_events.backend='db'. run_events.backend is '{run_events_backend}'. Set scheduler.multi_instance=false or configure run_events.backend: db.","messagePattern":"scheduler\\.multi_instance=true requires run_events\\.backend='db'\\. run_events\\.backend is '(.+?)'\\. Set scheduler\\.multi_instance=false or configure run_events\\.backend: db\\.","errorType":"console","errorClass":"SystemExit","httpStatus":null,"severity":"critical","filePath":"backend/app/gateway/deps.py","lineNumber":98,"sourceCode":"    immediately.\n    \"\"\"\n    try:\n        workers = int(os.environ.get(\"GATEWAY_WORKERS\", \"1\"))\n    except (TypeError, ValueError):\n        workers = 1\n\n    scheduler = getattr(config, \"scheduler\", None)\n    multi_instance_requested = bool(getattr(scheduler, \"multi_instance\", False))\n    multi_instance_scheduler = bool(getattr(scheduler, \"enabled\", False) and multi_instance_requested)\n\n    backend = getattr(config.database, \"backend\", None)\n    run_events_backend = getattr(getattr(config, \"run_events\", None), \"backend\", None)\n    run_ownership = getattr(config, \"run_ownership\", None)\n\n    if multi_instance_requested and backend != \"postgres\":\n        raise SystemExit(f\"scheduler.multi_instance=true requires database.backend='postgres'. database.backend is '{backend}'. Set scheduler.multi_instance=false or configure Postgres.\")\n    if multi_instance_requested and run_events_backend != \"db\":\n        raise SystemExit(f\"scheduler.multi_instance=true requires run_events.backend='db'. run_events.backend is '{run_events_backend}'. Set scheduler.multi_instance=false or configure run_events.backend: db.\")\n    if multi_instance_requested and (run_ownership is None or not run_ownership.heartbeat_enabled):\n        raise SystemExit(\"scheduler.multi_instance=true requires run_ownership.heartbeat_enabled=true so peer runs retain a valid lease. Set scheduler.multi_instance=false or enable run ownership heartbeats.\")\n\n    if workers <= 1:\n        return\n\n    if config.scheduler.enabled and not multi_instance_scheduler:\n        raise SystemExit(f\"GATEWAY_WORKERS={workers} cannot run with scheduler.enabled=true because each worker starts its own scheduler. Set GATEWAY_WORKERS=1, scheduler.multi_instance=true, or scheduler.enabled=false.\")\n\n    if _browser_tools_enabled_in_config(config):\n        raise SystemExit(browser_multi_worker_error(workers))\n\n    if backend != \"postgres\":\n        raise SystemExit(f\"GATEWAY_WORKERS={workers} requires database.backend='postgres', but database.backend is '{backend}'. SQLite cannot support concurrent multi-process access. Set GATEWAY_WORKERS=1 or switch to Postgres.\")\n\n    if run_events_backend != \"db\":\n        raise SystemExit(\n            f\"GATEWAY_WORKERS={workers} requires run_events.backend='db', but run_events.backend is '{run_events_backend}'. \"","sourceCodeStart":80,"sourceCodeEnd":116,"githubUrl":"https://github.com/bytedance/deer-flow/blob/1dd6ba1acb03700589994b0366c5d1c7d05e2eff/backend/app/gateway/deps.py#L80-L116","documentation":"SystemExit raised at startup when scheduler.multi_instance=true but run_events.backend is not 'db'. With multiple scheduler instances, run lifecycle events must live in the shared database; memory/JSONL stores are process-local, so peer schedulers could not observe each other's scheduled-run events and coordination guarantees would silently break.","triggerScenarios":"config.yaml with scheduler.multi_instance: true and run_events.backend set to memory or jsonl (or left at its default), for any database backend.","commonSituations":"Enabling multi_instance on a config that never configured run_events; default run_events backend differing from db in the example config; environment drift between staging and production config.","solutions":["Set run_events.backend: db in config.yaml","Or set scheduler.multi_instance: false","Restart the Gateway and confirm startup succeeds"],"exampleFix":"# config.yaml\n# before\nrun_events:\n  backend: memory\n# after\nrun_events:\n  backend: db","handlingStrategy":"validation","validationCode":"cfg = load_config()\nmi = bool(getattr(getattr(cfg, \"scheduler\", None), \"multi_instance\", False))\nre_backend = getattr(getattr(cfg, \"run_events\", None), \"backend\", None)\nassert not (mi and re_backend != \"db\"), \"multi_instance needs run_events.backend: db\"","typeGuard":null,"tryCatchPattern":null,"preventionTips":["When enabling multi_instance, set run_events.backend: db in the same commit","Diff config.yaml against config.example.yaml after upgrades — new required sections appear","Validate config in a pre-deploy step so bad config fails in CI, not at container boot"],"tags":["config","startup","scheduler","run-events","system-exit","gateway"],"backgroundTag":null,"analyzedSha":"1dd6ba1acb03700589994b0366c5d1c7d05e2eff","analyzedAt":"2026-08-14T21:20:34.804Z","schemaVersion":2},"datasetVersion":"2026-08-15T22:17:37.221Z"}