Diagnostic Settings Create API Has Inconsistent Behavior On logAnalyticsDestinationType for Certain Resource

[magodo]

The field logAnalyticsDestinationType in the diagnostic setting Creation API is not consistent and have skew for certain resources.

1. For some certain resource (e.g. azure app service), if the user calls diagnostic settings with logAnalyticsDestinationType set to Dedicated, the API returns successful, but the body of a following GET returns that field with null. Is it possible to raise an error if the resource doesn't support the specified destination type?
2. For some other resource (e.g. azure API management), if the user calls diagnostic settings with logAnalyticsDestinationType absent, the API returns successful, but the body of a following GET returns that field with AzureDiagnostics, which is not documented. Is it possible to return nil as documented?

Reference

• azurerm_monitor_diagnostic_setting re applies every time the log_analytics_destination_type property on plan/apply hashicorp/terraform-provider-azurerm#6676.

[ghost]

Thanks for the feedback! We are routing this to the appropriate team for follow-up. cc @antcp, @AzureAppServiceCLI.

[eissko]

Any progress on this here?

In my terraform state is always stored empty string of logAnalyticsDestinationType = "" for resource type azurerm_monitor_diagnostic_setting. It results in the change on terraform plan each time it is run.

using terraform 0.12.29
azurerm provider ">= 2.22.0"

related problems/issues:

hashicorp/terraform-provider-azurerm#6769
hashicorp/terraform-provider-azurerm#6676

[eissko]

FYI
related issue:
hashicorp/terraform-provider-azurerm#8131

[favoretti]

@ArcturusZhang ping on this one?

[ArcturusZhang]

This issue will need service team's attention to resolve.
Ping @antcp, @AzureAppServiceCLI

[bm77525-kr]

Any updates here @ArcturusZhang?

[aaronswainonset]

Any update on this?

[SenorRagequit]

Is anyone even working on this module?
This issue has been brought up many times and is still not fixed, just like the bug with the re-deploying of diagnostic settings, as the metrics get set to 0 instead of NULL.

[magodo]

Ping @akning-ms, can we priotirize this issue as it is keeping impacting users, for a long while.

[ritik-gupta-bzy]

any updates on this issue? Still, the behavior is the same for some resources (for example: when setting up azurerm_monitor_diagnostic_setting setting using terraform)