mirror of
https://github.com/Growstuff/growstuff.git
synced 2026-10-31 19:29:26 -04:00
Brenda Wallace
8995037a86
Puma was running in cluster mode with one worker, which keeps a master process holding a full copy of the preloaded app next to the worker. Heroku counts both, so the 512 MB dyno sat at ~620 MB with a constant R14. WEB_CONCURRENCY of 0 or 1 now runs one process with no master; two or more workers still use cluster mode with preload_app!. Puma reads WEB_CONCURRENCY itself, so single mode is set explicitly with `workers 0`. Added a comment explaining the memory budget.
62 lines
2.6 KiB
Ruby
62 lines
2.6 KiB
Ruby
# frozen_string_literal: true
|
|
|
|
# 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 }
|
|
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 }
|
|
|
|
# Specifies the `environment` that Puma will run in.
|
|
#
|
|
environment ENV.fetch("RAILS_ENV") { "development" }
|
|
|
|
# Specifies the `pidfile` that Puma will use.
|
|
pidfile ENV.fetch("PIDFILE") { "tmp/pids/server.pid" }
|
|
|
|
# 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).
|
|
#
|
|
# MEMORY BUDGET: production is a 512 MB Heroku Basic dyno. Heroku raises R14 at
|
|
# 512 MB and kills the dyno at ~1 GB (R15), so aim to keep the whole dyno below
|
|
# 500 MB. Heroku sums the RSS of every process in the dyno, and each Ruby
|
|
# process running this app costs roughly 200-300 MB, so the number of
|
|
# processes is the main lever:
|
|
#
|
|
# * Clustered mode with a single worker runs a master that holds a full copy
|
|
# of the preloaded app next to the worker, and both are counted. That was
|
|
# ~620 MB, permanently over quota.
|
|
# * So 0 or 1 means single mode: one process, no master. Concurrency comes
|
|
# from the threads (RAILS_MAX_THREADS), which are cheap compared to
|
|
# processes.
|
|
# * Two or more workers will not fit in 512 MB. Only raise WEB_CONCURRENCY
|
|
# above 1 after moving to a bigger dyno, and check `heroku logs` for R14
|
|
# afterwards.
|
|
#
|
|
web_concurrency = Integer(ENV.fetch("WEB_CONCURRENCY") { 0 })
|
|
|
|
# Puma reads WEB_CONCURRENCY itself, so single mode has to be set explicitly.
|
|
workers(web_concurrency > 1 ? web_concurrency : 0)
|
|
|
|
if web_concurrency > 1
|
|
# 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!
|
|
end
|
|
|
|
# Allow puma to be restarted by `rails restart` command.
|
|
plugin :tmp_restart
|