Trouble configuring rails passenger server w/ nginx

Hello, I need some help configuring raven-ruby (v 2.6.3) for a Rails 5 API-only application.

I cannot send any anything to sentry that involves an error in a controller action.
Also, have told raven to capture POST data.

For now, I have set up a controller, to manually trigger errors, then check the presence in sentry.

Sentry config file (config/initializers/sentry.rb):

Raven.configure do |config|
  config.environments = %w[staging production]
  config.sanitize_fields =
  config.rails_activesupport_breadcrumbs = true
  config.should_capture = { |_e| true } # capture everything

  config.transport_failure_callback = lambda do |event|
    Rails.logger.debug('---- Sentry Raven is broken')

  config.http_adapter = Faraday.default_adapter

  config.processors -= [Raven::Processor::PostData]

I get this message in my nginx error log:

INFO -- sentry: ** [Raven] Raven 2.6.3 configured not to capture errors: No host specified, no public_key specified, no secret_key specified, no project_id specified

But when using the rails console for the app, I can see that the Raven.configuration object clearly has all of those things.

I should also mention that I manually trigger errors from the console with
Raven.capture_exception('<err msg>') and by running the task, raven:test[<dsn>].
For what it’s worth, the app is also configured to use syslog and lograge,
but I don’t think those things matter because I can run the app locally and with the appropriate rails environment, I can trigger errors from controller actions.

What am I doing wrong?
Does it matter that I am using the same DSN for staging and production versions of the application?

Update, Solved

The problem was my understanding of the deployment environment. The server is passenger, which was reading environment variables from the nginx configuration.
It was not reading the environment variables from the shell, which is what I was expecting.

Going to leave this up here in case someone runs into a similar issue.