# This configuration file will be evaluated by Puma. The top-level methods that # are invoked here are part of Puma's configuration DSL. For more information # about methods provided by the DSL, see https://puma.io/puma/Puma/DSL.html. # # Puma starts a configurable number of processes (workers) and each process # serves each request in a thread from an internal thread pool. # # You can control the number of workers using ENV["WEB_CONCURRENCY"]. You # should only set this value when you want to run 2 or more workers. The # default is already 1. # # The ideal number of threads per worker depends both on how much time the # application spends waiting for IO operations and on how much you wish to # prioritize throughput over latency. # # As a rule of thumb, increasing the number of threads will increase how much # traffic a given process can handle (throughput), but due to CRuby's # Global VM Lock (GVL) it has diminishing returns and will degrade the # response time (latency) of the application. # # The default is set to 3 threads as it's deemed a decent compromise between # throughput and latency for the average Rails application. # # Any libraries that use a connection pool or another resource pool should # be configured to provide at least as many connections as the number of # threads. This includes Active Record's `pool` parameter in `database.yml`. # Calculate available memory for Ruby process sizing (in MB) available_memory_mb = if RUBY_PLATFORM =~ /darwin/ # Default to a reasonable amount on macOS where memory detection is difficult 8192 # 8GB default else begin # Try to get memory from /proc/meminfo on Linux mem_total_kb = `grep MemTotal /proc/meminfo`.to_s.strip.split(/\s+/)[1].to_i mem_total_mb = mem_total_kb / 1024 # If we couldn't detect memory, use a safe default mem_total_mb > 0 ? mem_total_mb : 4096 rescue 4096 # 4GB fallback if detection fails end end # Calculate workers based on available CPUs and memory # Each worker uses ~300-500MB of RAM, so we scale based on available memory # If WEB_CONCURRENCY is set, use that value instead cpu_count = begin require 'etc' Etc.nprocessors rescue 2 # Default to 2 if we can't detect end # Default worker calculation: # - With ample memory: Use CPU count # - With limited memory: Use memory-based calculation with a min of 2 and max based on CPUs # This automatically adapts to the environment default_workers = if ENV["RAILS_ENV"] == "development" 1 # Use 1 worker in development for simplicity else # Calculate based on available memory, assuming ~400MB per worker memory_based_workers = (available_memory_mb / 400).to_i # Ensure at least 2 workers for production (for zero-downtime restarts) # and cap at CPU count to avoid excessive context switching [memory_based_workers, 2].max end default_workers = [default_workers, cpu_count].min workers_count = ENV.fetch("WEB_CONCURRENCY") { default_workers } # Configure thread count for optimal throughput # More threads = better for IO-bound applications # Fewer threads = better for CPU-bound applications min_threads_count = ENV.fetch("RAILS_MIN_THREADS", 5) max_threads_count = ENV.fetch("RAILS_MAX_THREADS", 12) threads min_threads_count, max_threads_count # Specifies the `port` that Puma will listen on to receive requests; default is 3000. port ENV.fetch("PORT", 3000) # Allow puma to be restarted by `bin/rails restart` command. plugin :tmp_restart # Run the Solid Queue supervisor inside of Puma for single-server deployments # Enable by default in development for convenience, can be disabled with SOLID_QUEUE_IN_PUMA=false if (Rails.env.development? && ENV["SOLID_QUEUE_IN_PUMA"] != "false") || ENV["SOLID_QUEUE_IN_PUMA"] == "true" plugin :solid_queue end # Specify the PID file. Defaults to tmp/pids/server.pid in development. # In other environments, only set the PID file if requested. pidfile ENV["PIDFILE"] if ENV["PIDFILE"] # Set reasonable timeouts - these prevent hanging requests from consuming resources worker_timeout 60 worker_boot_timeout 60 # Preload the application to reduce memory footprint in production preload_app! if ENV["RAILS_ENV"] == "production" # When using preload_app, ensure that connections are properly handled on_worker_boot do ActiveRecord::Base.establish_connection if defined?(ActiveRecord) end # Log the configuration puts "Puma starting with #{workers_count} worker(s), #{min_threads_count}-#{max_threads_count} threads per worker" puts "Available system resources: #{cpu_count} CPU(s), #{available_memory_mb}MB RAM"