Create pipeline for restarts and hups that can be initiated by chatops
We can easily adapt the patcher to create a pipeline for hup'ing and restarting services. This would support:
- draining nodes from haproxy before the operation
- running the hup's and restarts in parallel across the different fleet types
- initiating it from chatops
The command, "service", or whatever we decide to call it would look something like this:
/chatops service --env <env> [--role <filter>] [--restart] [--hup]
Where the defaults would be to hup all services across all roles in a pipeline. The filters can be passed in to target roles like sidekiq, api, web, canary, etc.