In one project you said tagging releases faile you during an incident. How can tags causeproblems?

A short cautionary account of how manually tagging releases after deployment (rather than from the exact build commit) caused a mismatch between the git tag and what was actually running in production. During an incident, pulling the tag returned code that didn't match production, making debugging useless. The fix is to automate tagging in CI from the build commit, ensuring a single source of truth.