Inconsistent Error Handling Across Regions

“az vm create: RuntimeError('already consumed') masks the real deployment error (SkuNotAvailable/QuotaExceeded) ### Describe the bug When `az vm create` fails a deployment (e.g. `SkuNotAvailable` due to capacity restrictions, or `QuotaExceeded`), the CLI's own error-handling path throws a *second*, unrelated exception while trying to format the first one, so the top-level failure the user sees is a confusing internal traceback (`RuntimeError: The content for this response was already consumed`) instead of a clean error message. The real Azure error is still present deeper in the traceback text, but it's buried under CLI-internal noise. ### Related command ``` az vm create \ --resource-group <rg> \ --name <vm> \ --image Win2022Datacenter \ --size Standard_B2s \ --admin-username <user> \ --admin-password <redacted> \ --public-ip-address "" \ --storage-sku StandardSSD_LRS ``` This reproduced consistently across multiple regions (westeurope, northeurope, eastus) and multiple VM sizes (Standard_B2s, Standard_B2ms, Standard_D2s_v3, Standard_D2s_v5, Standard_B1s) whenever the underlying deployment failed - both for `SkuNotAvailable` (capacity restriction) and `QuotaExceeded` failures. ### Errors ``` Traceback (most recent call last): File "/opt/az/lib/python3.14/site-packages/azure/cli/core/commands/arm.py", line 109, in handle_template_based_exception raise CLIError(ex.inner_exception.error.message) ^^^^^^^^^^^^^^^^^^^^^^^^ AttributeError: 'NoneType' object has no attribute 'error' During handling of the above exception, another exception occurred: Traceback (most recent call last): File "/opt/az/lib/python3.14/site-packages/knack/cli.py", line 233, in invoke cmd_result = self.invocation.execute(args) File "/opt/az/lib/python3.14/site-packages/azure/cli/core/commands/__init__.py", line 677, in execute raise ex File "/opt/az/lib/python3.14/site-packages/azure/cli/core/commands/__init__.py", line 820, in _run_jobs_serially results.append(self._run_job(expanded_arg, cmd_copy)) File "/opt/az/lib/python3.14/site-packages/azure/cli/core/commands/__init__.py", line 812, in _run_job return cmd_copy.exception_handler(ex) File "/opt/az/lib/python3.14/site-packages/azure/cli/core/commands/arm.py", line 112, in handle_template_based_exception raise_subdivision_deployment_error(ex.response.internal_response.text, ex.error.code if ex.error else None) File "/opt/az/lib/python3.14/site-packages/requests/models.py", line 928, in text if not self.content: File "/opt/az/lib/python3.14/site-packages/requests/models.py", line 899, in content raise RuntimeError("The content for this response was already consumed") RuntimeError: The content for this response was already consumed ``` The genuine deployment error (e.g.): ``` (SkuNotAvailable) The requested VM size for resource 'Following SKUs have failed for Capacity Restrictions: Standard_B2s' is currently not available in location 'eastus'. Please try another size or deploy to a different location or different zone. ``` or ``` Code: QuotaExceeded Message: Operation could not be completed as it results in exceeding approved standardDSv5Family Cores quota. ... Current Limit: 0, Current Usage: 0, Additional Required: 2 ... ``` ...*is* present in the exception text at that point, but `raise_subdivision_deployment_error` tries to read `ex.response.internal_response.text` a second time after `requests` has already consumed the stream, so formatting it throws `RuntimeError` instead of surfacing the message. ### Expected behavior A failed deployment (capacity, quota, or otherwise) should surface a single, clean error message from the actual Azure response - not a `RuntimeError` about an already-consumed HTTP response body masking it. ### Environment summary ``` { "azure-cli": "2.89.1", "azure-cli-core": "2.89.1", "azure-cli-telemetry": "1.1.0" } ``` - OS: Linux (Ubuntu-based), Python 3.14 - Reproduced with b”