This is a question for anyone in the community to answer if capable.
Our company is planning on incorporating something along the lines of this: https://github.com/ryanpq/do-wpc into Trellis so that we can deploy to a scalable server infrastructure.
Was this type of automation being considered and if we are to implement this ourselves would you be interested in us sharing our findings once complete? Lastly, does anyone have any tips or bits of insight that might be useful for us as we venture towards this goal? Is there a starting point that we should be building on top of? Etc…
I’ve been fiddling with Salt-Cloud, a (somewhat) new addition to SaltStack. It has perfect integration with Digital Ocean, and with minimal setup I was able to initiate simple commands via command line to my Digital Ocean droplet.
I wrote about some basic findings here:
This could be used as a starting point for Trellis, or to run additional software or commands.
Depending on your end goal and time at your disposal you may want to keep in mind that you will probably need to script your own solution for autoscaling and that Digital Ocean do not have floating IP’s (in the case that you wanted HA). I was interested in trying to build HA sites with Digital Ocean and may still look at building ‘half’ of it ie due to no floating IP and including my own autoscaling. There is also talk of floating Ip’s being available some time this year (you could also look at round robin DNS however this can have its own issues). Anyway I am looking at it from a HA perspective ie for building and scaling MVP’s however for non HA sites they are already great in my opinion.
EDIT: I would also be setting up DB replication for the ‘half’ HA site which that doesn’t include in its deployment.