diff --git a/config/puma.rb b/config/puma.rb
index 56136f025..1c012b193 100644
--- a/config/puma.rb
+++ b/config/puma.rb
@@ -1,50 +1,99 @@
 # This file contains configuration settings for the Puma web server. These
-# settings apply when running Danbooru with `bin/rails server`. In production,
-# Danbooru currently uses Unicorn instead of Puma.
+# settings apply when running Danbooru with `bin/rails server`.
+#
+# The following environment variables are used:
+#
+# * RAILS_ENV
+# * PUMA_PORT
+# * PUMA_BIND
+# * PUMA_WORKERS
+# * PUMA_MIN_THREADS
+# * PUMA_MAX_THREADS
+# * PUMA_WORKER_TIMEOUT
+# * PUMA_PIDFILE
+# * PUMA_CONTROL_URL
+#
+# Use `bin/pumactl` to control a running Puma instance.
 #
 # @see https://puma.io
 # @see https://github.com/puma/puma
+# @see https://github.com/puma/puma/blob/319f84db13ee59f7b24885cec686d5c714998a4c/lib/puma/configuration.rb#L188 (default options)
+# @see https://devcenter.heroku.com/articles/deploying-rails-applications-with-the-puma-web-server
 
-# Puma can serve each request in a thread from an internal thread pool.
-# The `threads` method setting takes two numbers: a minimum and maximum.
-# Any libraries that use thread pools should be configured to match
-# the maximum value specified for Puma. Default is set to 5 threads for minimum
-# and maximum; this matches the default thread size of Active Record.
-#
-max_threads_count = ENV.fetch("RAILS_MAX_THREADS") { 5 }
-min_threads_count = ENV.fetch("RAILS_MIN_THREADS") { max_threads_count }
+# The server port or listening address. Default is http://0.0.0.0:3000.
+if ENV.has_key?("PUMA_PORT")
+  port ENV["PUMA_PORT"]
+elsif ENV.has_key?("PUMA_BIND")
+  bind ENV["PUMA_BIND"]
+else
+  # low_latency=true means TCP_NODELAY
+  # backlog=1024 means socket listen backlog
+  # https://github.com/puma/puma/blob/319f84db13ee59f7b24885cec686d5c714998a4c/lib/puma/dsl.rb#L193
+  bind "tcp://0.0.0.0:3000?low_latency=true&backlog=1024"
+end
+
+# The number of `workers` to boot in clustered mode. Workers are forked web
+# server processes. If using threads and workers together, the concurrency of
+# the application would be max `threads` * `workers`. Workers do not work on
+# JRuby or Windows (both of which do not support processes). The Postgres
+# connection limit may need to be raised for high `thread` * `worker` counts.
+if ENV.has_key?("PUMA_WORKERS")
+  workers ENV["PUMA_WORKERS"]
+elsif ENV["RAILS_ENV"] == "production"
+  workers ENV.fetch("PUMA_WORKERS", Etc.nprocessors)
+else
+  # Use single worker mode in development for easier debugging
+  workers 0
+end
+
+# The number of threads per worker to use. The `threads` method
+# setting takes two numbers: a minimum and maximum. Any libraries that use
+# thread pools should be configured to match the maximum value specified for
+# Puma.
+max_threads_count = ENV.fetch("PUMA_MAX_THREADS", 5)
+min_threads_count = ENV.fetch("PUMA_MIN_THREADS", 1)
 threads min_threads_count, max_threads_count
 
-# Specifies the `worker_timeout` threshold that Puma will use to wait before
-# terminating a worker in development environments.
-#
-worker_timeout 3600 if ENV.fetch("RAILS_ENV", "development") == "development"
+# Verifies that all workers have checked in to the master process within
+# the given timeout. If not the worker process will be restarted. This is
+# not a request timeout, it is to protect against a hung or dead process.
+# Setting this value will not protect against slow requests.
+worker_timeout ENV.fetch("PUMA_WORKER_TIMEOUT", 60)
 
-# Specifies the `port` that Puma will listen on to receive requests; default is 3000.
-#
-port ENV.fetch("PORT") { 3000 }
+# The number of seconds to wait for another request within a persistent (keep
+# alive) session.
+persistent_timeout 20
 
-# Specifies the `environment` that Puma will run in.
-#
-environment ENV.fetch("RAILS_ENV") { "development" }
+# The number of seconds to wait until we get the first data for the request
+first_data_timeout 30
 
-# Specifies the `pidfile` that Puma will use.
-pidfile ENV.fetch("PIDFILE") { "tmp/pids/server.pid" }
+# The `environment` that Puma will run in.
+environment ENV.fetch("RAILS_ENV", "development")
 
-# Specifies the number of `workers` to boot in clustered mode.
-# Workers are forked web server processes. If using threads and workers together
-# the concurrency of the application would be max `threads` * `workers`.
-# Workers do not work on JRuby or Windows (both of which do not support
-# processes).
-#
-# workers ENV.fetch("WEB_CONCURRENCY") { 2 }
+# The `pidfile` that Puma will use.
+pidfile ENV.fetch("PUMA_PIDFILE", "tmp/pids/server.pid")
 
 # Use the `preload_app!` method when specifying a `workers` number.
 # This directive tells Puma to first boot the application and load code
 # before forking the application. This takes advantage of Copy On Write
 # process behavior so workers use less memory.
-#
-# preload_app!
+if ENV["RAILS_ENV"] == "production"
+  preload_app!
+end
 
 # Allow puma to be restarted by `rails restart` command.
 plugin :tmp_restart
+
+# Start the Puma control rack application on +url+. This application can
+# be communicated with to control the main server. Additionally, you can
+# provide an authentication token, so all requests to the control server
+# will need to include that token as a query parameter. This allows for
+# simple authentication.
+#
+# Usage:
+#
+# * bin/pumactl stats
+# * bin/pumactl -C tcp://localhost:9293 stats
+#
+# https://github.com/puma/puma#controlstatus-server
+activate_control_app ENV.fetch("PUMA_CONTROL_URL", "tcp://localhost:9293"), no_token: true