✨ Add nodeSelector to procs

[RocKhalil]

Added the possibility of having node_name aka nodeName in the proc configuration.
This allows users to push each proc to a different node

[RocKhalil]

Updated to use nodeSelector with key and name;
example: node_pool_key: 'doks.digitalocean.com/node-pool', node_pool_name: 'api'

usage:

proc :web, 'bundle exec puma -C config/puma.rb', scale: 5, node_pool_key: 'doks.digitalocean.com/node-pool', node_pool_name: 'api'

[collimarco]

Probably it would be better to have something like node: { key: 'val', another_key: 'value' }

[RocKhalil]

@collimarco I can be an option, however it will still require custom key names to map directly into the kube file. so we'll need to do:

node: { pool_key: '..', pool_name: '...' }

[collimarco]

@RocKhalil No, it would be node: { 'pool_key': 'pool_name' }. This also allows to filter nodes based on other conditions, not just the pool name.

[RocKhalil]

@collimarco basically the pool_key has different values based on the provider.
example:
GKE: cloud.google.com/gke-nodepool
Digital Ocean: doks.digitalocean.com/node-pool

[collimarco]

@RocKhalil Yes, I know, it's what I said... node: { 'doks.digitalocean.com/node-pool': 'my-pool-1' }