Rebuild deployment procedure
All the evidence is showing that our deployment procedure is basically broken.
We tried to move to push jobs, this actually showed to be painful and actually made evident that we are greatly depending on the output of the task to debug and push the deployment process forward.
Our team is in discussions to get to a better deployment process that is both safe and effective.
We will use this issue to explain how will it be implemented and how will it work.
For the time being, we are moving away from using chef push jobs as they have proven to be highly unreliable.