Addition of Production Deployment Documentation
tpitale opened this issue · comments
It would be very nice to have a brief, theoretical explanation of how one might deploy Puma into production.
Questions that might be answered:
- How is it different to run in production than Unicorn or Thin or xyz?
- Is it appropriate to have Puma running behind Nginx or Apache, using a socket? Is this ideal?
- Thoughts on configuration of Puma to handle scale with threading? # of threads as compared to # of processes.
- JRuby or Rubinius vs. MRI
The docs could also:
- Link to Rails docs for threaded config (if such a thing exists?)
- Aaron Patterson's discussion of the threadsafe flag in Rails and process vs. threaded servers
- Other resources or suggested reading?
It may be beyond the scope of puma.io to support documentation on running any threaded server.
Rails 4 will have threadsafe on (or off, whatever the opposite is) by default, so that will help...
+1