"azuread_users" for huge group (~30K) stops after 20 minutes

[AlexeyKTS]

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritise this request
• Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do not help prioritise the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment

Terraform (and AzureAD Provider) Version

Terraform 1.5.5
azuread v2.41.0

Affected Resource(s)

azuread_users

Terraform Configuration Files

#Read users from the AAD source group
data "azuread_users" "users" {
  ignore_missing = true
  object_ids = data.azuread_group.source_group.members
  
  timeouts {
    read = "3h"
  }
}

Debug Output

UUID removed:

2023-08-21T21:07:23.1370968Z 2023-08-21T21:07:23.135Z [INFO]  provider.terraform-provider-azuread_v2.41.0_x5: 2023/08/21 21:07:23 [DEBUG] AzureAD Response: 200 OK for <...> (GET https://graph.microsoft.com/beta/users/<userID>): timestamp=2023-08-21T21:07:23.135Z
2023-08-21T21:07:23.1375118Z 2023-08-21T21:07:23.135Z [ERROR] provider.terraform-provider-azuread_v2.41.0_x5: Response contains error diagnostic: diagnostic_summary="Retrieving user with object ID: "<userID>"" tf_data_source_type=azuread_users tf_req_id=<...> @module=sdk.proto diagnostic_severity=ERROR tf_provider_addr=provider tf_rpc=ReadDataSource @caller=github.com/hashicorp/terraform-plugin-go@v0.14.3/tfprotov5/internal/diag/diagnostics.go:55 diagnostic_detail="UsersClient.BaseClient.Get(): could not read response body: context deadline exceeded" tf_proto_version=5.3 timestamp=2023-08-21T21:07:23.135Z
2023-08-21T21:07:23.1377514Z 2023-08-21T21:07:23.136Z [ERROR] vertex "data.azuread_users.users" error: Retrieving user with object ID: "<userID>"
2023-08-21T21:07:23.1378931Z 2023-08-21T21:07:23.136Z [ERROR] vertex "data.azuread_users.users (expand)" error: Retrieving user with object ID: "<userID>"
2023-08-21T21:07:23.1380514Z 2023-08-21T21:07:23.136Z [DEBUG] created provider logger: level=debug

Panic Output

Expected Behavior

There is a group about 30K users. Expected behavior is to read all of them.

Actual Behavior

After 20mins pipeline stops and shows next error:

2023-08-21T21:07:23.5394882Z �[31m│�[0m �[0m�[0m  with data.azuread_users.users,
2023-08-21T21:07:23.5395517Z �[31m│�[0m �[0m  on main.tf line 13, in data "azuread_users" "users":
2023-08-21T21:07:23.5395952Z �[31m│�[0m �[0m  13: data "azuread_users" "users" �[4m{�[0m�[0m
2023-08-21T21:07:23.5396303Z �[31m│�[0m �[0m
2023-08-21T21:07:23.5396725Z �[31m│�[0m �[0mUsersClient.BaseClient.Get(): could not read response body: context
2023-08-21T21:07:23.5397164Z �[31m│�[0m �[0mdeadline exceeded

Steps to Reproduce

1. Create AAD group with 19K users and more
2. Use azuread_users
3. terraform plan

Important Factoids

• Huge AAD group of users (~30k).
• Always 20 mins and stops.
• Number of read users always different, but between 18-19k.
• If ignore_missing = false then error a bit different:

  2023-08-21T19:53:12.8209316Z 2023-08-21T19:53:12.819Z [ERROR] provider.terraform-provider-azuread_v2.41.0_x5: Response contains error diagnostic: tf_proto_version=5.3 tf_req_id=<...> tf_rpc=ReadDataSource @caller=github.com/hashicorp/terraform-plugin-go@v0.14.3/tfprotov5/internal/diag/diagnostics.go:55 @module=sdk.proto diagnostic_detail="UsersClient.BaseClient.Get(): Get "https://graph.microsoft.com/beta/users/<userID>": http: RoundTripper implementation (*retryablehttp.RoundTripper) returned a nil *Response with a nil error" diagnostic_severity=ERROR diagnostic_summary="Retrieving user with object ID: "<userID>"" tf_data_source_type=azuread_users tf_provider_addr=provider timestamp=2023-08-21T19:53:12.819Z
  2023-08-21T19:53:12.8215756Z 2023-08-21T19:53:12.819Z [ERROR] vertex "data.azuread_users.users" error: Retrieving user with object ID: "<userID>"
  2023-08-21T19:53:12.8217218Z 2023-08-21T19:53:12.820Z [ERROR] vertex "data.azuread_users.users (expand)" error: Retrieving user with object ID: "<userID>"
  

• azuread_users does not react on any value in "timeouts" 5 min or 3 hours always giving 20 mins for run.

References

• #0000