|
2 | 2 |
|
3 | 3 | ## question |
4 | 4 | Hi there, I am trying to try to use the (Azure Monitor OTEL Exporter)[https://github.com/Azure/azure-sdk-for-python/tree/main/sdk/monitor/azure-monitor-opentelemetry-exporter#microsoft-opentelemetry-exporter-for-azure-monitor] for a python application, and I am having trouble getting data flow when I follow the examples of setting up the exporter. I tried to copy the (hello world sample code)[https://github.com/Azure/azure-sdk-for-python/tree/main/sdk/monitor/azure-monitor-opentelemetry-exporter#export-hello-world-trace] and put my connection string, and I'm still not seeing any tracing data land in my traces table in my App Insights instance. Could I get some help here to figure out what is going on, or where the sample code might not be functional? |
5 | | - |
| 5 | + |
6 | 6 | I tried to look through the debug logs: |
7 | 7 | ``` |
8 | 8 | 2025-05-20 22:57:57,387 DEBUG urllib3.connectionpool Starting new HTTP connection (1): 169.254.169.254:80 |
@@ -228,3 +228,100 @@ So. For you. I would recommend the following to unblock yourself. |
228 | 228 | ``` |
229 | 229 |
|
230 | 230 | Once you successfully record new recordings, test-proxy push and your Tag will be properly populated. |
| 231 | + |
| 232 | +# Question Regarding Regression in Python SDK v11.0.0 for Container Registry |
| 233 | + |
| 234 | +## question |
| 235 | +Hi Language - Python, I'm from the Azure Container Registry team, and we have a question regarding one of the recent SDK releases. |
| 236 | + |
| 237 | +The release in question is v11.0.0 ([AutoRelease] t2-containerregistry-2024-12-10-60943(can only be merged by SDK owner) by azure-sdk · Pull Request #38810 · Azure/azure-sdk-for-python), which was generated based on this Swagger PR Savaradh containerregistry microsoft.container registry 2024 11 01 preview new by savaradh · Pull Request #31612 · Azure/azure-rest-api-specs |
| 238 | + |
| 239 | +We recently noticed a regression in version 11.0.0: Previously, the _create_initial method returned a deserialized Task object, but in v11.0.0, it now returns a streamed response. This change was not mentioned in the changelog. |
| 240 | + |
| 241 | +Additionally, our Swagger PR did not modify the 2019-06-01-preview/containerregistry_build.json file, so we're unsure why the SDK PR includes changes to these operation files. |
| 242 | + |
| 243 | +In the "removed" lines (lines 503-507 in red), there was code that did: |
| 244 | +``` |
| 245 | +if response.status_code == 200: |
| 246 | + deserialized = self._deserialize("Task", pipeline_response) |
| 247 | +if response.status_code == 201: |
| 248 | + deserialized = self._deserialize("Task", pipeline_response) |
| 249 | +``` |
| 250 | +But in the current version (line 496 in green), it's doing: |
| 251 | +``` |
| 252 | +deserialized = response.stream_download(self._client._pipeline, decompress=_decompress) |
| 253 | +``` |
| 254 | + |
| 255 | +Could someone help us understand the root cause of this issue? Any insights would be greatly appreciated. Thanks |
| 256 | + |
| 257 | +## answer |
| 258 | +This was a change made in the code generator a while ago, and it's basically an implementation detail change in the public generated code. `_create_initial` is a private method that is used to set up the initial call to start the polling in `begin_create`. We just switched over our initial calls to not do premature deserialization, and there is 0 customer impact. Hope this clears stuff up! |
| 259 | + |
| 260 | +# OOB Stable Python SDK |
| 261 | + |
| 262 | +## question |
| 263 | +Hi Language - Python team, I'm trying to release Python SDK here: https://github.com/Azure/sdk-release-request/issues/6124 which is required to test and release Azure CLI module. Given that CLI has a deadline of next week for code complete, I have a couple of questions around this: |
| 264 | + |
| 265 | +1. Can this SDK be released out of band ASAP? |
| 266 | +2. For future iterations, what's the process we should take to align with CLI releases, so that we have enough time to release SDKs and then release CLI as well? |
| 267 | +Thanks |
| 268 | + |
| 269 | +## answer |
| 270 | +1. SDK team will release this package in advance to unblock CLI. Add Chenxi Jiang (WICRESOFT NORTH AMERICA LTD) for awareness. |
| 271 | +2. If CLI module depends on SDK, it is better to make release request issue earlier so that SDK team could prepare the release in advance. |
| 272 | + |
| 273 | +# GA review |
| 274 | + |
| 275 | +## question |
| 276 | +Hi all, we're looking to release our first stable Python SDK in (Deid GA release by jovinson-ms · Pull Request #40850 · Azure/azure-sdk-for-python)[https://github.com/Azure/azure-sdk-for-python/pull/40850]. Would appreciate a look at the APIView, as we pulled in some suggested client customization from Java and .NET reviews but would like feedback on whether they are idiomatic to Python. |
| 277 | + |
| 278 | +## answer |
| 279 | +Hi, I would recommend making a post in the (Language - Python - Reviews)[https://teams.microsoft.com/l/channel/19%3A4175567f1e154a80ab5b88cbd22ea92f%40thread.skype/Language%20-%20Python%20-%20Reviews?groupId=3e17dcb0-4257-4a30-b843-77f47f1d4121&tenantId=72f988bf-86f1-41af-91ab-2d7cd011db47] channel |
| 280 | + |
| 281 | +# Changelog release requirements |
| 282 | + |
| 283 | +## question |
| 284 | +Hi all, I'm trying to release our data plane SDK's first stable version in (Deid GA release by jovinson-ms · Pull Request #40850 · Azure/azure-sdk-for-python)[https://github.com/Azure/azure-sdk-for-python/pull/40850]. I'm getting (build failures on changelog validation)[https://dev.azure.com/azure-sdk/public/_build/results?buildId=4824427&view=logs&j=b70e5e73-bbb6-5567-0939-8415943fadb9&t=ac8f4042-9b76-5db4-27b1-2a4abaa9bb3c&l=101]. Looking through the docs, I don't see any changelog update process for data plane in contrast to the control plane automation. Can someone help me understand what the process is here? |
| 285 | + |
| 286 | +## answer |
| 287 | +You didn't update the base version of the package! Appears to still be `1.0.0b1` |
| 288 | +I believe it should be `1.0.0` to match the changelog you added |
| 289 | + |
| 290 | +# CSpell config |
| 291 | + |
| 292 | +## question |
| 293 | +Hi all, I know there are common CSpell config settings as .vscode/cspell.json. I also see a few package-level cspell config files checked in: (Code search results)[https://github.com/search?q=repo%3AAzure%2Fazure-sdk-for-python+path%3Acspell.json&type=code]. Is there some config that needs to be set for pipelines to use local cspell config? |
| 294 | + |
| 295 | +## answer |
| 296 | +We haven't done this in the SDK repos yet but we intend to allow folks to inherit from the root file. Similar to https://github.com/Azure/azure-rest-api-specs/blob/main/specification/communication/cspell.yaml where it imports the root cspell config. So as long as you inherit it should work. |
| 297 | + |
| 298 | +# Disable Pylint checks on .pyi files |
| 299 | + |
| 300 | +## question |
| 301 | +Hi all, |
| 302 | + |
| 303 | +Storage is adding .pyi files for each of our clients to move `kwargs` to named keywords without having to make massive changes to our runtime code. We have noticed that pylint still runs on these .pyi files and a lot of the checks just don't make sense there, for example `super-init-not-called`. Obviously super init would not be called in a stub. |
| 304 | + |
| 305 | +Example PR: ([Storage] [Named Keywords] [Blob] `_container_client.pyi` and aio by weirongw23-msft · Pull Request #41030 · Azure/azure-sdk-for-python)[https://github.com/Azure/azure-sdk-for-python/pull/41030/files] |
| 306 | + |
| 307 | +This is causing us to have to put in a bunch of pylint disables. Curious on everyone's thoughts on simply disabling pylint on stub files entirely? |
| 308 | + |
| 309 | +## answer |
| 310 | +You could consider using `Unpack` with `TypedDict` to handle the typing of kwargs? In the long term it would probably be a lower maintenance solution (it also allows the addition of documentation in intellisense prompts - which the current pyi file doesn't seem to have). |
| 311 | +Here's an example if you want to have a read through: |
| 312 | +(azure-sdk-for-python/sdk/projects/azure-projects/azure/projects/resources/storage/_resource.py at main · Azure/azure-sdk-for-python)[https://github.com/Azure/azure-sdk-for-python/blob/main/sdk/projects/azure-projects/azure/projects/resources/storage/_resource.py#L237-L243] |
| 313 | + |
| 314 | +# Python Storage SDK - `walk_blobs` API |
| 315 | + |
| 316 | +## question |
| 317 | +Hi Anna and Kashif Khan, |
| 318 | + |
| 319 | +(A customer mentioned)[https://github.com/Azure/azure-sdk-for-python/issues/40873] that `walk_blobs` API could return `BlobPrefix` where the current rtype is just `BlobProperties`. Based on this feedback, it appears that `walk_blob` needs to have rtype `BlobProperties | BlobPrefix` ((PR)[https://github.com/Azure/azure-sdk-for-python/pull/40931]). However, this would be a breaking change to our customers, and it would be harder to work with as the customer needs to type check themselves, so not great. |
| 320 | + |
| 321 | +In .NET, `GetBlobsByHierarchy` (returns a custom type)[https://github.com/Azure/azure-sdk-for-net/blob/6ee19685b3c1ecc7f9d7b6318954b12708dcc179/sdk/storage/Azure.Storage.Blobs/src/BlobContainerClient.cs#L2713-L2773] of `Pageable<BlobHierarchyItem>` which has a Boolean value to check whether the item is `BlobPrefix` or `BlobProperties`. |
| 322 | + |
| 323 | +We'd like some feedback on how to best proceed here. Thanks! |
| 324 | + |
| 325 | +## answer |
| 326 | +To clarify, changing the type hint is not a breaking change, just annoying to work with a Union return type. The other options of introducing a wrapper class however, would be breaking obviously. |
| 327 | +If this is already acting that way at runtime, it means people already use `isintance` or `try/except AttributeError` anyway, so the `Union` still make things better. |
0 commit comments