[FEATURE] retries implemented, or max_retries configable if retry already implemented

[FireballDWF]

Is your feature request related to a problem?

"elastic: Error 503 (Service Unavailable)" is sometimes returned when the access policy creation which allows usage has not achieved eventually consistency before invocation of resource creation like opensearch_index which actually depends on the policy.

What solution would you like?

retries on 503 errors, and other transient failures.

What alternatives have you considered?

could add sleep between the two calls

Do you have any additional context?

Implement retries similar to the AWS provider behavior - https://registry.terraform.io/providers/hashicorp/aws/latest/docs#max_retries

[rblcoder]

In aws.Config, MaxRetries is currently set to 1

terraform-provider-opensearch/provider/provider.go

Line 458 in 3e84564

MaxRetries: aws.Int(1),