[core-lro] "The long-running operation has failed" is an unclear message for customers

[kazrael2119]

when sdk operation call was failing we were getting this info in the error response "The long-running operation has failed" and this message is not useful. Instead of it if we could get exact reason in the error message it will be very helpful.

"The long-running operation has failed" is an unclear message when we get this error. We can't know where the error comes from so that we don't know how to fix it.

Could we improve the experience for it?

[deyaaeldeen]

The design guidelines say there should be an error field that contains an error if the LRO status is Failed, see bullet 5 in here. I think we can update core-lro to print the content of that error field in this case. What do you think?

[kazrael2119]

The design guidelines say there should be an error field that contains an error if the LRO status is Failed, see bullet 5 in here. I think we can update core-lro to print the content of that error field in this case. What do you think?

Sounds great

[MaryGao]

@deyaaeldeen Agreed. In our recent experience, when the provisioningState is failed the core-lro would throw the exception directly. If we'd like to know what the exact error message is we have to dig deep into the response body via onResponse callback or debug into detailed payload. The more useful information would be the error code and message.

One thing needs more attention here is we will meet the LRO failure error both in two stages, one is in polling process the other is in final get. Could we improve both to get these errors?

Case1 - Payload throws during polling

{
  "id": "/subscriptions/xxx/providers/Microsoft.Workloads/locations/EASTUS/operationStatuses/xxx",
  "name": "xxx",
  "resourceId": "/subscriptions/xxx/resourceGroups/sdk-testing-vidya6_22_3_2023/providers/Microsoft.Workloads/monitors/haMonitor4/providerInstances/promethueusHa-provider",
  "status": "Failed",
  "startTime": "2023-03-23T10:49:28.6416273Z",
  "error": {
    "code": "InvalidProviderNameLength",
    "message": "\\n      The length of given Provider name: 'promethueusHa-provider' is not valid.\\n      Recommended Action: The valid length of provider name is minimum of 2 and maximum of 20 characters long.\\n    "
  }
}

Case 2: Payload during final get

{
  "id": "/subscriptions/xxx/resourceGroups/sdk-testing-monitor_23_3_2023/providers/Microsoft.Workloads/monitors/haMonitor",
  "name": "haMonitor",
  "type": "microsoft.workloads/monitors",
  "systemData": {
    "createdBy": "5910ab76-9a08-40f9-88d4-48d5f01393a8",
    "createdByType": "Application",
    "createdAt": "2023-03-23T10:01:26.039Z",
    "lastModifiedBy": "5910ab76-9a08-40f9-88d4-48d5f01393a8",
    "lastModifiedByType": "Application",
    "lastModifiedAt": "2023-03-23T10:05:29.264Z"
  },
  "tags": { "sdk-testing": "true" },
  "location": "eastus",
  "provisioningState": "Failed",
  "errors": {
    "code": "ManagedResourceGroupAlreadyExists",
    "message": "\n      Managed Resource Group creation failed.\n      Possible Causes: Resource Group with Provided name already exists and managed by resource with Id '/subscriptions/49d64d54-e966-4c46-a868-1999802b762c/resourceGroups/sdk-testing-monitor_21_3_2023/providers/Microsoft.Workloads/monitors/haMonitor'.\n      Recommended Action: Retry the operation with another Managed Resource Group Name. If the issue persists, please create an issue here :- https://aka.ms/AAgc4fy.\n    ",
    "details": []
  },
...


}

/cc @gargankit-microsoft

[suhanishri]

Hi @MaryGao , @gargankit-microsoft, I tried the addition of options parameter to the client start/stop requests and I'm able to see the error information in the raw logging. Attaching the output snapshot tested on "Stop Central Server Instance" test and the provisioning failed

Creating RG : sdk-testing-vis6_JS_29_3_2023
Raw Response: {
status: 200,
headers: HttpHeadersImpl {
_headersMap: Map(15) {
'cache-control' => [Object],
'pragma' => [Object],
'transfer-encoding' => [Object],
'content-type' => [Object],
'content-encoding' => [Object],
'expires' => [Object],
'etag' => [Object],
'vary' => [Object],
'x-ms-ratelimit-remaining-subscription-reads' => [Object],
'x-ms-request-id' => [Object],
'x-ms-correlation-request-id' => [Object],
'x-ms-routing-request-id' => [Object],
'strict-transport-security' => [Object],
'x-content-type-options' => [Object],
'date' => [Object]
}
},
request: PipelineRequestImpl {
url: 'https://management.azure.com/subscriptions/49d64d54-e966-4c46-a868-1999802b762c/providers/Microsoft.Workloads/locations/EASTUS2EUAP/operationStatuses/4b38211b-1372-4a45-ab1a-fc4996c4aec6*8AEAFF279C9AF8A9037951D6A3A2A1D6507221CC592E1E361EABEC79A23726FE?api-version=2023-04-01',
body: undefined,
headers: HttpHeadersImpl {
_headersMap: [Map]
},
method: 'GET',
timeout: 0,
formData: undefined,
disableKeepAlive: false,
proxySettings: undefined,
streamResponseStatusCodes: Set(0) {},
withCredentials: false,
abortSignal: undefined,
tracingOptions: undefined,
onUploadProgress: undefined,
onDownloadProgress: undefined,
requestId: '190064e3-9f06-4177-9c8e-9e2adeb1cd0c',
allowInsecureConnection: false,
enableBrowserStreams: false
},
bodyAsText: '{"id":"/subscriptions/49d64d54-e966-4c46-a868-1999802b762c/providers/Microsoft.Workloads/locations/EASTUS2EUAP/operationStatuses/4b38211b-1372-4a45-ab1a-fc4996c4aec68AEAFF279C9AF8A9037951D6A3A2A1D6507221CC592E1E361EABEC79A23726FE","name":"4b38211b-1372-4a45-ab1a-fc4996c4aec68AEAFF279C9AF8A9037951D6A3A2A1D6507221CC592E1E361EABEC79A23726FE","resourceId":"/subscriptions/49d64d54-e966-4c46-a868-1999802b762c/resourceGroups/DemoRGVIS/providers/Microsoft.Workloads/sapVirtualInstances/DRT/centralInstances/cs0","status":"Failed","startTime":"2023-03-29T09:05:42.2190468Z","endTime":"2023-03-29T09:05:44.2440293Z","error":{"code":"SapOpsValidationFailureUnsupportedState","message":"\n\t The Stop operation failed on the CentralInstance cs0.\n\t Possible Causes: The SAP resource is already in Unavailable.\n\t Recommended Action: The CentralInstance should be in Running or PartiallyRunning status for the Stop oepration to begin. The current status is Unavailable.\n "}}',
parsedBody: {
id: '/subscriptions/49d64d54-e966-4c46-a868-1999802b762c/providers/Microsoft.Workloads/locations/EASTUS2EUAP/operationStatuses/4b38211b-1372-4a45-ab1a-fc4996c4aec68AEAFF279C9AF8A9037951D6A3A2A1D6507221CC592E1E361EABEC79A23726FE',
name: '4b38211b-1372-4a45-ab1a-fc4996c4aec68AEAFF279C9AF8A9037951D6A3A2A1D6507221CC592E1E361EABEC79A23726FE',
status: 'Failed',
startTime: 2023 - 03 - 29 T09: 05: 42.219 Z,
endTime: 2023 - 03 - 29 T09: 05: 44.244 Z,
error: {
code: 'SapOpsValidationFailureUnsupportedState',
message: '\n' +
'\t The Stop operation failed on the CentralInstance cs0.\n' +
'\t Possible Causes: The SAP resource is already in Unavailable.\n' +
'\t Recommended Action: The CentralInstance should be in Running or PartiallyRunning status for the Stop oepration to begin. The current status is Unavailable.\n' +
' '
},
resourceId: '/subscriptions/49d64d54-e966-4c46-a868-1999802b762c/resourceGroups/DemoRGVIS/providers/Microsoft.Workloads/sapVirtualInstances/DRT/centralInstances/cs0'
}
}

We are getting proper error message in the field
"error":{"code":"SapOpsValidationFailureUnsupportedState","message":"\n\t The Stop operation failed on the CentralInstance cs0.\n\t Possible Causes: The SAP resource is already in Unavailable.\n\t Recommended Action: The CentralInstance should be in Running or PartiallyRunning status for the Stop oepration to begin. The current status is Unavailable.\n "}}'

[gargankit-microsoft]

1. @suhanishri seems like a client error, should we retry with correct payload here
2. @suhanishri error message seems to have typos, should we raise issue with service team?
3. @MaryGao , seems similar to what we already logged in ams RT too.

[MaryGao]

@gargankit-microsoft Offline confirmed with Suhani, I think this is the similar issue so we will track in this issue also.

[deyaaeldeen]

I merged in a fix and I am planning on making a release in May. Do we need to release sooner?

[MaryGao]

Thanks for the fix, I think we can release it in May.

[MaryGao]

@kazrael2119 Could you help bump the version of core-lro in codegen?

[deyaaeldeen]

@kazrael2119 Could you help bump the version of core-lro in codegen?

Please note this fix has not been released yet.

[kazrael2119]

@kazrael2119 Could you help bump the version of core-lro in codegen?

Please note this fix has not been released yet.

I'll keep track of this

[MaryGao]

Close it as fixed.