Error with RUM flush failed.

[lanzmlt]

What happened?

Hello, I'm trying to run qwen-code completely locally using qwen3-30b in a completely isolated environment.

~/.qwen/settings.json

{
  "security": {
    "auth": {
      "selectedType": "openai",
      "apiKey": "\"\"",
      "baseUrl": "https://llm-api-test.local/api/v1"
    }
  },
  "model": {
    "name": "qwen3-30B"
  },
  "telemetry": {
    "enabled": false
  },
  "enableOpenAILogging": false,
  "usageStatisticsEnabled": false,
  "$version": 2
}

Env in my system:

printenv

OPENAI_BASE_URL=https://llm-api-test.local/api/v1
OPENAI_MODEL=qwen3-30B

~/.qwen/.env

DEBUG=true
NODE_EXTRA_CA_CERTS=/usr/local/share/ca-certificates/server-cert.crt
HTTP_PROXY=http://proxy-server.local:3333
HTTPS_PROXY=http://proxy-server.local:3333
QWEN_TELEMETRY_DISABLED=true
QWEN_DISABLE_RUM=true
QWEN_ANALYTICS_DISABLED=true
DISABLE_TELEMETRY=true
NO_PROXY=*
QWEN_NO_STREAM=true

The request to the model API via curl is successful:

curl

❯ curl --location 'https://llm-api-test.local/api/v1/chat/completions' \ 
  --header 'Content-Type: application/json' \
  --data '{
    "messages": [
      {
        "role": "user",
        "content": "How can you help me?"
      }
    ],
    "model": "qwen3-30B"
  }'
{"id":"chatcmpl-93c39fff-7369-485a-919a-646f14ca8116","object":"chat.completion","created":1766594759,"model":"qwen3-30b-a3b-instruct-2507-fp8","choices":[{"index":0,"message":{"role":"assistant","reasoning_content":null,"content":"I can help you with a wide range of tasks! Here are some examples:
......
I’m here to help. 😊","tool_calls":[]},"logprobs":null,"finish_reason":"stop","stop_reason":null}],"usage":{"prompt_tokens":14,"total_tokens":266,"completion_tokens":252,"prompt_tokens_details":null},"prompt_logprobs":null,"kv_transfer_params":null}%

When i try to use qwen-code i get errors:

qwen "Analyze QWEN.md"

❯ qwen "Analyze QWEN.md" \
  --chat-recording=false \
  --max-session-turns 1 \
  --telemetry=false \
  --exclude-tools task list_directory read_file grep_search glob read_many_files save_memory todo_write exit_plan_mode web_fetch
[DEBUG] CLI: Delegating hierarchical memory load to server for CWD: /home/user/qwen-cli/qwen-test (memoryImportFormat: tree)
[DEBUG] [MemoryDiscovery] Loading server hierarchical memory for CWD: /home/user/qwen-cli/qwen-test (importFormat: tree)
[DEBUG] [MemoryDiscovery] Searching for QWEN.md starting from CWD: /home/user/qwen-cli/qwen-test
[DEBUG] [MemoryDiscovery] Determined project root: None
[DEBUG] [BfsFileSearch] Scanning [1/200]: batch of 1
[DEBUG] [MemoryDiscovery] Final ordered QWEN.md paths to read: ["/home/user/qwen-cli/qwen-test/QWEN.md"]
[DEBUG] [MemoryDiscovery] Successfully read and processed imports: /home/user/qwen-cli/qwen-test/QWEN.md (Length: 0)
[DEBUG] [MemoryDiscovery] Combined instructions length: 0
ToolRegistry created []
Flushing log events to RUM.
Session ID: ae82c407-82d1-4891-b0f3-709fd8f1ef23
RUM flush failed. Error: Request failed with status 403
    at ClientRequest.<anonymous> (file:///home/user/.nvm/versions/node/v22.21.1/lib/node_modules/@qwen-code/qwen-code/cli.js:79442:29)
    at Object.onceWrapper (node:events:634:26)
    at ClientRequest.emit (node:events:519:28)
    at HTTPParser.parserOnIncomingClient (node:_http_client:772:27)
    at HTTPParser.parserOnHeadersComplete (node:_http_common:122:17)
    at Socket.socketOnData (node:_http_client:614:22)
    at Socket.emit (node:events:519:28)
    at Readable.read (node:internal/streams/readable:782:10)
    at Socket.read (node:net:785:39)
    at flow (node:internal/streams/readable:1283:53)
QwenLogger: Re-queued 1 events for retry (queue size: 2)
Flushing log events to RUM.
RUM flush failed. Error: Request failed with status 403
    at ClientRequest.<anonymous> (file:///home/user/.nvm/versions/node/v22.21.1/lib/node_modules/@qwen-code/qwen-code/cli.js:79442:29)
    at Object.onceWrapper (node:events:634:26)
    at ClientRequest.emit (node:events:519:28)
    at HTTPParser.parserOnIncomingClient (node:_http_client:772:27)
    at HTTPParser.parserOnHeadersComplete (node:_http_common:122:17)
    at Socket.socketOnData (node:_http_client:614:22)
    at Socket.emit (node:events:519:28)
    at Readable.read (node:internal/streams/readable:782:10)
    at Socket.read (node:net:785:39)
    at flow (node:internal/streams/readable:1283:53)
QwenLogger: Re-queued 1 events for retry (queue size: 1)
Flushing log events to RUM.
OpenAI API Streaming Error: Connection error.
Error when talking to API Full report available at: /tmp/qwen-client-error-Turn.run-sendMessageStream-2025-12-24T16-32-56-785Z.json
[API Error: Connection error.]

❯ cat /tmp/qwen-client-error-Turn.run-sendMessageStream-2025-12-24T16-32-56-785Z.json
{
  "error": {
    "message": "Connection error.",
    "stack": "Error: Connection error.\n    at OpenAI.makeRequest (file:///home/user/.nvm/versions/node/v22.21.1/lib/node_modules/@qwen-code/qwen-code/cli.js:137917:17)\n    at process.processTicksAndRejections (node:internal/process/task_queues:105:5)\n    at async file:///home/user/.nvm/versions/node/v22.21.1/lib/node_modules/@qwen-code/qwen-code/cli.js:140685:27\n    at async ContentGenerationPipeline.executeWithErrorHandling (file:///home/user/.nvm/versions/node/v22.21.1/lib/node_modules/@qwen-code/qwen-code/cli.js:140825:26)\n    at async retryWithBackoff (file:///home/user/.nvm/versions/node/v22.21.1/lib/node_modules/@qwen-code/qwen-code/cli.js:142371:22)\n    at async GeminiChat.makeApiCallAndProcessStream (file:///home/user/.nvm/versions/node/v22.21.1/lib/node_modules/@qwen-code/qwen-code/cli.js:143473:33)\n    at async file:///home/user/.nvm/versions/node/v22.21.1/lib/node_modules/@qwen-code/qwen-code/cli.js:143430:33\n    at async Turn.run (file:///home/user/.nvm/versions/node/v22.21.1/lib/node_modules/@qwen-code/qwen-code/cli.js:145946:28)\n    at async GeminiClient.sendMessageStream (file:///home/user/.nvm/versions/node/v22.21.1/lib/node_modules/@qwen-code/qwen-code/cli.js:166003:26)\n    at async file:///home/user/.nvm/versions/node/v22.21.1/lib/node_modules/@qwen-code/qwen-code/cli.js:335437:26"
  },
  "context": [
    {
      "role": "user",
      "parts": [
        {
          "text": "This is the Qwen Code. We are setting up the context for our chat.\nToday's date is Wednesday, December 24, 2025 (formatted according to the user's locale).\nMy operating system is: linux\nI'm currently working in the directory: /home/user/qwen-cli/qwen-test\nHere is the folder structure of the current working directories:\n\nShowing up to 20 items (files + folders).\n\n/home/user/qwen-cli/qwen-test/\n└───QWEN.md"
        }
      ]
    },
    {
      "role": "model",
      "parts": [
        {
          "text": "Got it. Thanks for the context!"
        }
      ]
    },
    {
      "role": "user",
      "parts": [
        {
          "text": "Analyze QWEN.md"
        }
      ]
    },
    [
      "Analyze QWEN.md"
    ]
  ]
}%

What did you expect to happen?

Correct operation in a completely isolated environment with access only to the API AI model on another virtual machine

Client information

Client Information

Run qwen to enter the interactive CLI, then run the /about command.

❯ qwen /about
[DEBUG] CLI: Delegating hierarchical memory load to server for CWD: /home/user/qwen-cli/qwen-test (memoryImportFormat: tree)
[DEBUG] [MemoryDiscovery] Loading server hierarchical memory for CWD: /home/user/qwen-cli/qwen-test (importFormat: tree)
[DEBUG] [MemoryDiscovery] Searching for QWEN.md starting from CWD: /home/user/qwen-cli/qwen-test
[DEBUG] [MemoryDiscovery] Determined project root: None
[DEBUG] [BfsFileSearch] Scanning [1/200]: batch of 1
[DEBUG] [MemoryDiscovery] Final ordered QWEN.md paths to read: ["/home/user/qwen-cli/qwen-test/QWEN.md"]
[DEBUG] [MemoryDiscovery] Successfully read and processed imports: /home/user/qwen-cli/qwen-test/QWEN.md (Length: 0)
[DEBUG] [MemoryDiscovery] Combined instructions length: 0
ToolRegistry created [
  'task',
  'list_directory',
  'read_file',
  'grep_search',
  'glob',
  'read_many_files',
  'save_memory',
  'todo_write',
  'exit_plan_mode',
  'web_fetch'
]
Flushing log events to RUM.
Session ID: fb54ea6a-8756-434a-b7f4-11081ffcbae4
RUM flush failed. Error: Request failed with status 403
    at ClientRequest.<anonymous> (file:///home/user/.nvm/versions/node/v22.21.1/lib/node_modules/@qwen-code/qwen-code/cli.js:79442:29)
    at Object.onceWrapper (node:events:634:26)
    at ClientRequest.emit (node:events:519:28)
    at HTTPParser.parserOnIncomingClient (node:_http_client:772:27)
    at HTTPParser.parserOnHeadersComplete (node:_http_common:122:17)
    at Socket.socketOnData (node:_http_client:614:22)
    at Socket.emit (node:events:519:28)
    at Readable.read (node:internal/streams/readable:782:10)
    at Socket.read (node:net:785:39)
    at flow (node:internal/streams/readable:1283:53)
QwenLogger: Re-queued 1 events for retry (queue size: 2)
Flushing log events to RUM.
RUM flush failed. Error: Request failed with status 403
    at ClientRequest.<anonymous> (file:///home/user/.nvm/versions/node/v22.21.1/lib/node_modules/@qwen-code/qwen-code/cli.js:79442:29)
    at Object.onceWrapper (node:events:634:26)
    at ClientRequest.emit (node:events:519:28)
    at HTTPParser.parserOnIncomingClient (node:_http_client:772:27)
    at HTTPParser.parserOnHeadersComplete (node:_http_common:122:17)
    at Socket.socketOnData (node:_http_client:614:22)
    at Socket.emit (node:events:519:28)
    at Readable.read (node:internal/streams/readable:782:10)
    at Socket.read (node:net:785:39)
    at flow (node:internal/streams/readable:1283:53)
QwenLogger: Re-queued 1 events for retry (queue size: 1)

Login information

No response

Anything else we need to know?

No response

[tanzhenxin]

@lanzmlt Sorry for the inconvenience. You can disable usage statistics tracking by setting privacy.usageStatisticsEnabled to false. Please refer to the settings documentation for details: https://github.com/QwenLM/qwen-code/blob/main/docs/users/configuration/settings.md

[liqiongyu]

Opened PR #1427.

Root cause: if settings.json already has $version: 2 but still uses deprecated top-level keys (e.g. usageStatisticsEnabled), we skip migration and silently ignore the legacy key, so privacy.usageStatisticsEnabled can remain enabled and QwenLogger keeps trying to flush RUM (noisy RUM flush failed logs in isolated envs).

Fix: apply legacy-key aliases in-memory during settings loading (legacy key -> v2 path when missing; v2 nested value wins). Added regression tests for v2+legacy keys.