Explicitly avoid broken aws 2.65.0 terraform provider (#2688)

[sre-bot]

cherry-pick #2688 to release-1.1

---

What problem does this PR solve?

hashicorp/terraform-provider-aws#13626

root@3d1552defe47:/tidb-operator/deploy/aws# terraform apply
module.tidb-operator.module.eks.data.aws_caller_identity.current: Refreshing state...
module.tidb-operator.module.eks.data.aws_region.current: Refreshing state...
module.tidb-operator.module.eks.data.aws_iam_policy_document.workers_assume_role_policy: Refreshing state...
module.tidb-operator.module.eks.data.aws_iam_policy_document.cluster_assume_role_policy: Refreshing state...
module.tidb-operator.module.eks.data.aws_ami.eks_worker: Refreshing state...
module.vpc.data.aws_availability_zones.available: Refreshing state...
module.bastion.data.aws_ami.centos: Refreshing state...
Error: expected private_ip to contain a valid IPv4 address, got:
  on .terraform/modules/bastion.ec2/terraform-aws-ec2-instance-2.3.0/main.tf line 96, in resource "aws_instance" "this_t2":
  96: resource "aws_instance" "this_t2" {

What is changed and how does it work?

This patch simply tells terraform to avoid version 2.65.0 of the AWS provider.

Check List

Tests

• Manual test (add detailed scripts or steps below)

Deploy worked fine after making this change.

Code changes

• Has Terraform scripts change

I'm not sure if it's really a Terraform script change.

Side effects

Unknown.

Related changes

• Need to cherry-pick to the release branch
• Need to update the documentation

Does this PR introduce a user-facing change?:

NONE

[Yisaer]

/merge

[sre-bot]

Your auto merge job has been accepted, waiting for:

• 2689

[sre-bot]

/run-all-tests

[Yisaer]

/merge

[sre-bot]

/run-all-tests

[sre-bot]

@sre-bot merge failed.

[Yisaer]

/run-e2e-in-kind

[Yisaer]

/merge

[sre-bot]

/run-all-tests