[[["易于理解","easyToUnderstand","thumb-up"],["解决了我的问题","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["很难理解","hardToUnderstand","thumb-down"],["信息或示例代码不正确","incorrectInformationOrSampleCode","thumb-down"],["没有我需要的信息/示例","missingTheInformationSamplesINeed","thumb-down"],["翻译问题","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["最后更新时间 (UTC):2025-08-18。"],[],[],null,["# Known issues for Workflows\n\nThis page lists known issues for Workflows.\n\nYou can also check for existing issues or open new issues in the\n[public issue trackers](/support/docs/issue-trackers).\n\nPlacement of `for` directly after `try`\n---------------------------------------\n\nPlacing `for` directly after `try` causes an error. For example, a single step\ncan be placed directly after `try`, like this: \n\n```yaml\n- try:\n try:\n call: sys.log\n args:\n data: works\n retry: ${http.default_retry}\n```\n\nHowever, if you position `for` after the `try`, the step fails, and you cannot\ndeploy the workflow. For example: \n\n```yaml\n- try:\n try:\n for:\n value: v\n range: [1,2]\n steps:\n - log:\n call: sys.log\n args:\n data: ${v}\n retry: ${http.default_retry}\n```\n\nThe error message is as follows:\n\n`Could not deploy workflow: failed to build: error in step try: loop step name should not be empty (Code: 3)`\n\nThe workaround is to add a named step after the `try`. For example: \n\n```yaml\n- try:\n try:\n steps:\n - loopStep:\n for:\n value: v\n range: [1,2]\n steps:\n - log:\n call: sys.log\n args:\n data: ${v}\n retry: ${http.default_retry}\n```\n\nEvents larger than the maximum arguments size\n---------------------------------------------\n\nIf using Workflows as a destination for an\nEventarc trigger, events larger than the maximum\nWorkflows arguments size will fail to trigger workflow\nexecutions. For more information, see\n[Quotas and limits](/workflows/quotas).\n\n`HTTP request lost` message in logs\n-----------------------------------\n\nWhen executing a workflow that calls Cloud Build, your workflow fails and\nthere is an `HTTP request lost` message in the logs similar to the following: \n\n```bash\n[1500] HTTP request lost\nINTERNAL MESSAGE: HTTP request lost\n...\nCAUSED BY:\nRPC::UNREACHABLE: RPC connection timed out: FDD 20s, last read 2022-10-14 16:39:04 -0700 PDT\n```\n\nIf you encounter this error, try modifying your workflow by implementing either a\n[retry policy](/workflows/docs/reference/syntax/retrying) or through explicit\n[exception handling](/workflows/docs/reference/syntax/catching-errors#try-retry-except).\n\nCall logging and `accessString` method to retrieve secret data\n--------------------------------------------------------------\n\nIf the\n[call logging level](/workflows/docs/log-workflow#call_logging) is set to\n`log-all-calls` when\n[using the `accessString` helper method to retrieve secret data](/workflows/docs/reference/googleapis/secretmanager/v1/projects.secrets.versions/accessString),\nthe secret value is not redacted, and is printed in plain text to the logs in\n`jsonPayload.succeeded.response`.\n\nLong-running operation exception when using Cloud Resource Manager connector\n----------------------------------------------------------------------------\n\nThe Resource Manager connector method,\n[`googleapis.cloudresourcemanager.v3.projects.patch`](/workflows/docs/reference/googleapis/cloudresourcemanager/v3/projects/patch),\ndoesn't return a long-running operation (LRO) name. Even for a successful\nrequest, an exception similar to the following might be raised: \n\n```json\nexception: \"{\"message\":\"Long-running operation returned unexpected response.\",\n\"operation\":{\"done\":true,\"response\":{\"@type\":\"type.googleapis.com/google.cloud.resourcemanager.v3.Project\",\n...\n\"tags\":[\"ResponseTypeError\"]}\"\n```\n\nTo avoid an LRO polling error,\nset the `skip_polling` connector parameter to `true` so that the connector\ninvocation call is non-blocking if the initial request succeeds. A successful\nrequest returns `\"done\":true`; otherwise, to catch any exceptions use a\n[`try/except` structure](/workflows/docs/reference/syntax/catching-errors).\nFor more information, see the\n[Connectors reference](/workflows/docs/reference/googleapis).\n\nHTTP requests to Google Kubernetes Engine (GKE)\n-----------------------------------------------\n\nEvery GKE cluster has a control plane that handles Kubernetes API\nrequests. The control plane has two kinds of endpoints for cluster access:\n*DNS-based endpoints* and *IP-based endpoints* . For more information, see\n[Control plane access](/kubernetes-engine/docs/concepts/network-isolation#control-plane-access).\n\nWorkflows no longer supports\n[HTTP requests](/workflows/docs/http-requests) to the\n[IP-based endpoints](/kubernetes-engine/docs/reference/rest/v1/projects.locations.clusters#ipendpointsconfig)\nof GKE cluster control planes. For more information about the\nscope and impact of this deprecation, see the\n[service announcement](/workflows/docs/service-announcements#2025-01-28).\n\nTo ensure that your workflow functions as expected, you must access\n[DNS-based endpoints](/kubernetes-engine/docs/reference/rest/v1/projects.locations.clusters#dnsendpointconfig).\nTo access the DNS-based endpoints, we recommend that you use the Kubernetes API\nconnector in a workflow. For more information, see\n[Access Kubernetes API objects using a connector](/workflows/docs/access-kubernetes-api).\n\nWhat's next\n-----------\n\nLearn strategies helpful when [troubleshooting issues](/workflows/docs/troubleshooting)."]]