[Text Analytics] Implement IsEnvironmentReadyAsync

[joseharriaga]

Resolves: #35452.

Implemented IsEnvironmentReadyAsync with a simple request to the dynamic resource to confirm that permissions have propagated successfully, because otherwise we're at risk of having a bunch of test failures like this one:

Azure.RequestFailedException : Access denied due to invalid subscription key or wrong API endpoint. Make sure to provide a valid key for an active subscription and use a correct regional API endpoint for your resource.
Status: 401 (PermissionDenied)
ErrorCode: 401

Additionally, while testing, I found another instance of the transient Internal Server Error. The difference is that the message is simply "Internal Server Error." while the previous known case said "Failed to process task after several retry":

Azure.RequestFailedException : Internal Server Error.
Status: 200 (OK)
ErrorCode: InternalServerError

Additional Information:
Target: #/tasks/items/1
error-0: InternalServerError: Internal Server Error.

I added this one to the ShouldRetry method of RetryOnInternalServerErrorAttribute.

[azure-sdk]

API change check

API changes are not detected in this pull request.

[joseharriaga]

/azp run net - textanalytics - tests

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[joseharriaga]

/azp run net - textanalytics - tests

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[joseharriaga]

/azp run net - textanalytics - tests

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[joseharriaga]

/azp run net - textanalytics - tests

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[deyaaeldeen]

Looks good, however, the issue seems to be not TA specific. Should the pipeline add some latency after the resource creation step to mitigate it?

[heaths]

Looks good, however, the issue seems to be not TA specific. Should the pipeline add some latency after the resource creation step to mitigate it?

We don't want to add latency to all pipelines. Few, hopefully, need it. See my review for what I suggested to inject latency when needed.