Disable auto-scaling group for GitLab node

Presently there is an auto-scaling group to restart the GitLab app nodes in case of a failure. Unfortunately there is a critical issue here, namely that all repos reside on the attached EBS disk.

This means that if a node goes down and is unreachable, is terminated, that content could be lost. The new GitLab instance would then come up but without any repos.

Until we have a solution for a shared/common storage (EFS is not an option due to its performance characteristics) we should disable the auto scaling as people will want to (and need to) try to repair the current instance.