fix: handle race condition in release promote job

[AaronFeledy]

Problem

When a fresh (non-prerelease) release is created, GitHub fires both published and released events simultaneously. The promote job (triggered by released) races the deploy job (triggered by published) and tries to npm dist-tag add before the package is published to npm.

Fix

Add a retry loop that waits up to 5 minutes for the package version to appear on npm before promoting. If the version never appears, the job fails with a clear error instead of silently succeeding.

This is better than simply skipping when the version doesn't exist (as in lando/phpmyadmin#72), because it still catches genuine failures — e.g., if a prerelease promotion runs and the version is truly missing.

Affected repos

This same fix is being rolled out to all Lando repos with this release workflow pattern.

Closes any race condition failures on the promote job.

---

Note

Low Risk
Workflow-only change that adds a bounded wait/retry around npm dist-tag promotion; low blast radius but could cause release jobs to fail if npm propagation is unusually slow.

Overview
Prevents the promote release workflow job from racing the npm publish by adding a retry loop that waits (up to ~5 minutes) for the just-released version to become visible on npm before running npm dist-tag add ... latest.

If the version never appears, the job now fails with an explicit error instead of attempting the tag update immediately.

^{Written by Cursor Bugbot for commit a9852cb. This will update automatically on new commits. Configure here.}

[netlify]

❌ Deploy Preview for lando-platformsh failed. Why did it fail? →

Name	Link
🔨 Latest commit	a9852cb
🔍 Latest deploy log	https://app.netlify.com/projects/lando-platformsh/deploys/699b6d500b2f7700082bc2da