Can't change desired count for a task that didn't change within the last 3 months

[xalvarez]

Terraform and AWS Provider Version

Terraform v1.12.2
on linux_amd64
+ provider registry.terraform.io/hashicorp/aws v6.4.0

Affected Resource(s) or Data Source(s)

• aws_ecs_service

Expected Behavior

When changing the desired count of an ECS service, apply should work even if the corresponding ECS task is old.

Actual Behavior

Since AWS provider version 6.4.0, ListServiceDeployments is used to get the latest deployment without time filters. If no filter is provided, by default the API looks for a deployment from the last 3 months. If none is found, terraform apply fails.

Relevant Error/Panic Output

Error: waiting for ECS Service (arn:aws:ecs:eu-central-1:xxxxx) update: operation error ECS: ListServiceDeployments, https response error StatusCode: 400, RequestID: xxxxx, InvalidParameterException: Before and After time filter must be within the last 3 months.

Sample Terraform Configuration

Click to expand configuration

resource "aws_ecs_service" "service" {
  name                             = "example-service-dev"
  cluster                          = aws_ecs_cluster.service-cluster.id
  task_definition                  = aws_ecs_task_definition.task.arn
  desired_count                    = 1
  launch_type                      = "FARGATE"
  health_check_grace_period_seconds = 180
  wait_for_steady_state            = true

  network_configuration {
    security_groups  = [data.aws_security_group.alb_security_group.id, aws_security_group.service_security_group.id]
    subnets          = tolist(data.aws_subnets.private.ids)
    assign_public_ip = false
  }

  load_balancer {
    target_group_arn = aws_lb_target_group.target_group.id
    container_name   = "example-container"
    container_port   = 8080
  }

  lifecycle {
    create_before_destroy = true
  }
}

Steps to Reproduce

1. Change desired count
2. Run terraform apply

Debug Logging

Click to expand log output

http.request.body=
| {"cluster":"arn:aws:ecs:REGION:ACCOUNT_ID:cluster/SERVICE_NAME","createdAt":{"after":1736836063.319},"service":"SERVICE_NAME"}
net.peer.name=ecs.REGION.amazonaws.com rpc.method=ListServiceDeployments rpc.system=aws-api http.method=POST http.request.header.authorization="[REDACTED]" tf_resource_type=aws_ecs_service http.request.header.amz_sdk_invocation_id=[REDACTED] tf_mux_provider="*schema.GRPCProviderServer" http.request.header.x_amz_security_token="[REDACTED]" http.request_content_length=166 tf_aws.sdk=aws-sdk-go-v2 http.request.header.x_amz_date=[REDACTED] tf_rpc=ApplyResourceChange http.request.header.content_type=application/x-amz-json-1.1 http.user_agent="APN/1.0 HashiCorp/1.0 Terraform/1.12.2 (+https://www.terraform.io) terraform-provider-aws/6.4.0 (+https://registry.terraform.io/providers/hashicorp/aws) aws-sdk-go-v2/1.36.5 ua/2.1 os/linux lang/go#1.24.5 md/GOOS#linux md/GOARCH#amd64 api/ecs#1.60.0 m/r,t,u" rpc.service=ECS aws.region=REGION http.request.header.x_amz_target=AmazonEC2ContainerServiceV20141113.ListServiceDeployments http.url=https://ecs.REGION.amazonaws.com/ tf_provider_addr=registry.terraform.io/hashicorp/aws tf_req_id=[REDACTED] timestamp="2025-07-18T14:48:08.635+0200"

2025-07-18T14:48:08.700+0200 [DEBUG] provider.terraform-provider-aws_v6.4.0_x5: HTTP Response Received:
http.response.body=
| {"__type":"InvalidParameterException","message":"Before and After time filter must be within the last 3 months."}
http.response.header.date="Fri, 18 Jul 2025 12:48:08 GMT" rpc.method=ListServiceDeployments tf_aws.sdk=aws-sdk-go-v2 tf_aws.signing_region="" tf_provider_addr=registry.terraform.io/hashicorp/aws http.status_code=400 @caller=github.com/hashicorp/aws-sdk-go-base/v2@v2.0.0-beta.65/logging/tf_logger.go:45 http.duration=64 http.response.header.content_type=application/x-amz-json-1.1 http.response_content_length=113 rpc.system=aws-api tf_req_id=[REDACTED] tf_resource_type=aws_ecs_service @module=aws aws.region=REGION http.response.header.x_amzn_requestid=[REDACTED] rpc.service=ECS tf_mux_provider="*schema.GRPCProviderServer" tf_rpc=ApplyResourceChange timestamp="2025-07-18T14:48:08.700+0200"

2025-07-18T14:48:08.701+0200 [DEBUG] provider.terraform-provider-aws_v6.4.0_x5: request failed with unretryable error https response error StatusCode: 400, RequestID: [REDACTED], InvalidParameterException: Before and After time filter must be within the last 3 months.: @module=aws aws.region=REGION rpc.method=ListServiceDeployments tf_provider_addr=registry.terraform.io/hashicorp/aws tf_resource_type=aws_ecs_service tf_aws.sdk=aws-sdk-go-v2 rpc.system=aws-api @caller=github.com/hashicorp/aws-sdk-go-base/v2@v2.0.0-beta.65/logging/tf_logger.go:45 rpc.service=ECS tf_mux_provider="*schema.GRPCProviderServer" tf_req_id=[REDACTED] tf_rpc=ApplyResourceChange timestamp="2025-07-18T14:48:08.701+0200"

2025-07-18T14:48:08.705+0200 [ERROR] provider.terraform-provider-aws_v6.4.0_x5: Response contains error diagnostic: tf_rpc=ApplyResourceChange @module=sdk.proto diagnostic_severity=ERROR diagnostic_summary="waiting for ECS Service (arn:aws:ecs:REGION:ACCOUNT_ID:service/SERVICE_NAME/SERVICE_NAME) update: operation error ECS: ListServiceDeployments, https response error StatusCode: 400, RequestID: [REDACTED], InvalidParameterException: Before and After time filter must be within the last 3 months." tf_proto_version=5.9 tf_provider_addr=registry.terraform.io/hashicorp/aws tf_req_id=[REDACTED] @caller=github.com/hashicorp/terraform-plugin-go@v0.28.0/tfprotov5/internal/diag/diagnostics.go:58 diagnostic_detail="" tf_resource_type=aws_ecs_service timestamp="2025-07-18T14:48:08.704+0200"

2025-07-18T14:48:08.709+0200 [DEBUG] State storage *remote.State declined to persist a state snapshot

2025-07-18T14:48:08.709+0200 [ERROR] vertex "module.fargate-service.aws_ecs_service.service" error: waiting for ECS Service (arn:aws:ecs:REGION:ACCOUNT_ID:service/SERVICE_NAME/SERVICE_NAME) update: operation error ECS: ListServiceDeployments, https response error StatusCode: 400, RequestID: [REDACTED], InvalidParameterException: Before and After time filter must be within the last 3 months.

GenAI / LLM Assisted Development

n/a

Important Facts and References

No response

Would you like to implement a fix?

No

[github-actions]

Community Guidelines

This comment is added to every new Issue to provide quick reference to how the Terraform AWS Provider is maintained. Please review the information below, and thank you for contributing to the community that keeps the provider thriving! 🚀

Voting for Prioritization

• Please vote on this Issue by adding a 👍 reaction to the original post to help the community and maintainers prioritize it.
• Please see our prioritization guide for additional information on how the maintainers handle prioritization.
• Please do not leave +1 or other comments that do not add relevant new information or questions; they generate extra noise for others following the Issue and do not help prioritize the request.

Volunteering to Work on This Issue

• If you are interested in working on this issue, please leave a comment.
• If this would be your first contribution, please review the contribution guide.
• For new resources and data sources, use skaff to generate scaffolding with comments detailing common expectations.

[djglaser]

Hi, we will patch this bug and release ASAP. In the meantime, either an older version of the provider can temporarily be used, or wait_steady_state can be set to false. In the latter case, Terraform would trigger UpdateService and subsequently exit, without monitoring for stabilization. One would need to monitor the AWS console in order to ensure that stabilization works as expected.

[jar-b]

Relates #43434