How will the azurerm provider behave when API limits are reached?

I’m provisioning a large number for resources, and I’m looking at the -parallelism terraform parameter. I’ve successfully run with values of 20, 30, and 50.

However, I’m wondering what will happen when I meet or exceed the API limits documented here: Request limits and throttling - Azure Resource Manager | Microsoft Docs

Does the azurerm provider handle backing down gracefully? Will it honor the “Retry-After” responses, and… retry after x minutes?

I haven’t encountered an API limit yet, but expect I might sooner or later.


The go-autorest library used by the AzureRM provider implements retrying on 429 responses and uses the Retry-After header. This behavior gets updated from time to time. As of now the provider is using go-autorest v0.11.17 - you can refer to the changelog and code there for additional clarity.

Great! Thank you for your response.