Currently there's only one variable controlling storage directives for all worker types equally. The problem with this solution is that the needs of different worker nodes are not the same. For instance, the write workers typically require much more space than read workers.
This problem is particularly visible in our PS7 staging env (3 units of each worker type) - to have 100G volume attached to each write worker we need to allocate 900G of storage in total. This unnecessarily bubbles our quota.
We should allow setting storage directives separately for each worker type to optimize the deployment's resources footprint.
I started working on it in #182