There are basic 2 types of config in Elixir code:

1. Compile time configuration. Before Elixir 1.10 you could call Application.get_env/{2,3} in any part of the code, even compile time, so people often did stuff like:

   defmodule Foo do
     @option Application.get_env(:my_app, :foo_config)
   
     # …
   end
   

   This caused the @option to be set during compilation, so if someone was using releases (Distillery or Relx at the time) could be confused why @option didn’t changed when the configuration was changed on the target server. Since 1.10 we have Application.compile_env/{2,3} and Application.get_env/{2,3} will print warning when used in compile time. There are few examples of applications that use compile time env and in few situations this still sometimes makes some sense (for example poor-man dependency injection).
2. Runtime environment, that is Application.get_env/{2,3} calls within functions body, that is it. The “funny” thing there is that runtime configuration is also divided in 2 separate things (and people often forgot about it):
   1. “Startup config” that is read only at the beginning of application lifetime (often in module that is implementing Application behaviour, or in init/1 functions of the servers). This is for example reason why simple Application.put_env(:logger, :level, :warning) will not work, and you need to use Logger.configure(level: :warning) or reason why you cannot configure :kernel and :stdlib applications via config/config.exs (so for example configuration of Erlang’s logger is currently enormous PITA in Elixir).
   2. “On-demand” configuration where Application.get_env/{2,3} is called “in-place”, where it is needed. Reason why a lot components cannot be configured that way are mainly 2:
      • Performance - reading environment is much slower than just variable from the process state
      • Data is needed at the start of the process - for example you cannot change port on which Cowboy is listening for connections after it was started (obviously)

So 1. is obviously something that you cannot configure in runtime (examples of such behaviour is :mime database). And depending when your Vapor configuration will be ran 1.1. can also be off limits, at least for some applications (notably :kernel and :stdlib are hard to configure with any Elixir tooling).

The Application.env is 100% ets:

iex(serenity@wintermute.skunkwerks.at)9> :ets.tab2list :ac_tab
[
  {{:application_master, :credentials_obfuscation}, #PID<0.367.0>},
  {{:env, :gettext, :default_locale}, "en"},
  {{:env, :kernel, :logger},
   [
     {:handler, :default, :logger_std_h,
      %{
        config: %{type: :standard_io},
        formatter: {:logger_formatter,
         %{legacy_header: true, single_line: false}}
      }}
   ]},
  {{:application_master, :cowboy}, #PID<0.415.0>},
  {{:env, :amqp_client, :writer_gc_threshold}, 1000000000},
  {{:env, :lager, :crash_log_count}, 5},
  {{:env, :ex_unit, :timeout}, 60000},
  {{:application_master, :goldrush}, #PID<0.326.0>},
  {{:loaded, :logger},
...

You can look in the observer app, if you select view -> ETS tables -> system tables too.

For better (read-only) performance, look at Erlang -- persistent_term the persistent term cache, added in OTP21.2. Note that updating this is a java-esque stop-the-world scenario for the entire VM.

• Clever use of persistent_term – A Blog from the Erlang/OTP team – The Erlang/OTP team at Ericsson, the implementors and maintainers of Erlang/OTP.
• Use :persistent_term for Logger config by josevalim · Pull Request #8977 · elixir-lang/elixir · GitHub where Elixir starts using PT for Logger app

The first thing I can think off are modules names that macros will be used from, or will be modules generated by macros.

Basically, anything that changes the compilation result.

It is more like reading from ETS AFAIK. That is why it is not the best performance-wise when used in environment that need to be as fast as possible (like logger calls).

You cannot do stuff like:

config :kernel, :logger, [
  # …
]

In your config/config.exs nor in config/runtime.exs as when these files are evaluated by Mix kernel application is already started and running. This mean that you cannot easily configure logger in development that will work from the beginning of the VM lifetime (like for example SASL startup messages).

Let just say, that this isn’t very popular approach in Elixir development to use Erlang’s logger to full extent, but with Elixir 1.11 it can became more and more feasible to do so, as Elixir backends do not have simple access to structured logging (it will always be translated to plain string) nor to all log levels (these have access to “basics” - debug/info/warn/error, while Elixir 1.11 supports all 7 syslog levels). There are PRs to OTP to improve that, but it will take some time before that functionality will be available in Elixir.

Forgive me for my struggle to understand a good solution. I think my options are to do maybe these:

Option 1: fill runtime.exs:

if config_env() == :test do
  config :myapp, 
      x: "test-x",
      y: "test-y",
      z: System.get_env("TEST_Z", "test-z"),
     # ... etc...
else
  config :myapp, 
      x: System.get_env("X"),
      y: System.get_env("Y"),
      z: System.get_env("Z"),
     # ... etc...
end

Pro: declarative, clear
Con: double long runtime.exs. Also if-statements in configs are always for me a smell.

Option 2: use test_helper.exs

Use simple runtime.exs and add test values to test_helper.exs like maybe this:

  config :myapp, 
      x: System.get_env("X"),
      y: System.get_env("Y"),
      z: System.get_env("Z"),
     # ... etc...

and then in test_helper.exs I can done Application.put_env like

Application.put_env(:myapp, :x, "test-x")
Application.put_env(:myapp, :y, "test-y")
Application.put_env(:myapp, :z, "test-z")

Pro: simple runtime.exs (no double, no if-statement)
Con: cannot enforce System ENV vars using System.fetch_env! – only get_env would work so that execution does not stop before test_helper.exs runs and a value is supplied.

Option 3: Vapor

I looked also at Vapor. I like this: it could replace runtime.exs 100% I think. But it is still a problem when you must manually take a value and use Application.put_env to put it into the correct place so your app and deps can find it. For example, imagine a dependency with config like this:

config :some_dep, 
   deep: %{
      thing1: "a",
      thing2: System.get_env("THING2")
   }

With vapor, I can get THING2 ENV var from .env but I must be careful with any deep config like this example dependancy. Where to put it once I have it? I think I must merge map and then put inside carefully into Application process dictionary.

Pro: support for .env and .env.{ENV} files for each override. Errors when config is not correct.
Con: difficult for complex config that rely on Application.get_env.

I think there must be other ways. Sorry for my confusion. I admit I thought this was a easy topic but I am understanding more that config has subtle!