[Infra]: Refresh AWS token before each Terraform pipeline stage

[ClaireSimmonds]

Why is this issue important?

Because the "Deploy to Production" workflow includes a manual step to verify the terraform plan output, it is possible for the AWS token to expire before the "Terraform Apply" job kicks off, depending on how long the manual review takes. This may result in a failed build and deployers needing to re-run the pipeline to produce a successful build. This change will ensure that the token is generated at the beginning of each stage it is needed, which will reduce the need for additional manual effort and better accommodate longer build times in the future.

Current State

The pipeline currently generates the needed AWS credentials just before entering the "Terraform Plan" job. However, the plan needs to be reviewed manually before applying in production environments. This can result in a delay longer than the session time, depending on how long the review takes, and the key may not be valid by the time the "Terraform Apply" job executes.

Expected State

The pipeline will generate necessary AWS credentials for Terraform plan and apply at the beginning of each stage they are used.

Implementation Plan

Update the "Deploy to Production" workflow to:

1. Rename the current aws-auth job and associated references to aws-auth-plan
2. Create a new aws-auth-apply job that triggers before the tf-apply job