release-deploy.yml: deploy-to-prod passes cluster_name_staging instead of cluster_name_prod

[mawbid]

Bug

In release-deploy.yml, the deploy-to-prod job passes the wrong cluster name to deploy-ecs-service:

# line 142 — should be cluster_name_prod
CLUSTER_NAME: ${{ inputs.cluster_name_staging }}

CLUSTER_NAME is used by the aws ecs update-service --force-new-deployment and aws ecs wait services-stable steps inside deploy-ecs-service. Because FORCE_NEW_DEPLOYMENT defaults to false and is not passed in release-deploy.yml, these steps are currently skipped — so the bug is latent today. However, enabling FORCE_NEW_DEPLOYMENT on a release deployment would issue the force-new-deployment call against the staging cluster instead of prod.

Fix

Change line 142 from:

CLUSTER_NAME: ${{ inputs.cluster_name_staging }}

to:

CLUSTER_NAME: ${{ inputs.cluster_name_prod }}

A PR with this fix is open: https://github.com/aproorg/github-workflows/pull/$(gh pr create --repo aproorg/github-workflows --head mawbid:fix/prod-cluster-name --base main --title "fix: pass cluster_name_prod to deploy-to-prod job" --body "$(cat <<'PRBODY'
Fixes the deploy-to-prod job in release-deploy.yml which was passing cluster_name_staging as CLUSTER_NAME instead of cluster_name_prod.

CLUSTER_NAME is only used by the force-new-deployment and wait services-stable steps in deploy-ecs-service, and those steps are currently skipped because FORCE_NEW_DEPLOYMENT defaults to false. The bug is therefore latent — but enabling FORCE_NEW_DEPLOYMENT on a release deploy would target the staging cluster instead of prod.

One-line change: cluster_name_staging → cluster_name_prod on the CLUSTER_NAME input of the prod deployment step.
PRBODY
)" 2>&1 | grep -o '[0-9]*$')