Invalid formatting of logging messages when using slave nodes

I am using various Erlang slaves nodes for integration tests. They are running locally on my computer.

When using calls such as, the message do not use the formatting defined in config/config.ex. Instead they use some sort of default formatting, like this:

17:20:14.716 [info] Connecting to 'localhost' on 1234

The formatting is defined like this:

config :logger, :console,
  metadata: [:module, :pid],
  format: "$date $time $metadata[$level]$levelpad: $message\n"

I pushed this configuration on the slave node but it did not solve the problem. Any ideas?

Answering my own question… I was able to solve the problem by using a call to Logger.backend_configure on the remote node. Something similar to that:

rpc(node, Logger, :configure_backend,
      [:console, Application.fetch_env!(:logger, :console)])