Lowering *-prepare-disable-env job timeout

Context

In deployer pipeline, the *-prepare-disable-env jobs are used as mutex to make sure that there is always only one deployment happens to an environment at a time. Currently, these jobs have no timeout, so they use the 2h timeout from runners (example)

Previously, this long timeout makes sense, because auto-deploy pipelines start continuously for every auto-deploy package. However, with the recent split between build and deploy pipelines, we now do deploys smarter, by only deploying the next package once the previous package is done one gprd-cny (<implementation epic link needed>). Effectively, in normal circumstances, there is no overlap between deployments to the same environment. If it does, it must be something wrong.

In this issue, I suggest us to reduce the timeout of *-prepare-disable-env job to a reasonable number (e.g. 15min). This change speeds up the feedback loop, so release managers can react to issues faster, effectively reduce possible downtime and blocking time to GitLab.com.