hashi_uwsgi.ini

[uwsgi]
# Sources used to create this configuration file:
# https://uwsgi-docs.readthedocs.io/en/latest/Options.html
# https://uwsgi-docs.readthedocs.io/en/latest/ThingsToKnow.html
# https://www.reddit.com/r/Python/comments/4s40ge/understanding_uwsgi_threads_processes_and_gil/
# https://www.techatbloomberg.com/blog/configuring-uwsgi-production-deployment/
socket = /tmp/kolibri_hashi_uwsgi.sock
chmod-socket = 660
gid = www-data
#chown-socket = $(KOLIBRI_USER):www-data
chdir = /usr/lib/python3/dist-packages/
pythonpath = /usr/lib/python3/dist-packages/kolibri/dist
master = true    # https://uwsgi-docs.readthedocs.io/en/latest/Glossary.html?highlight=master
harakiri = 5000  # https://uwsgi-docs.readthedocs.io/en/latest/FAQ.html#what-is-harakiri-mode
enable-threads = false
cpu-affinity = 1    # Set the number of cores (CPUs) to allocate to each worker process
listen = 100        # Set the socket listen queue size. When this queue is full, requests will be rejected.
# rawrouter-buffer-size = 16392  # set internal buffer size
limit-as = 1024      # limit processes address space/vsz
reload-on-rss = 512 # reload if rss memory is higher than specified megabytes
no-orphans = true   # automatically kill workers if master dies
reload-mercy = 4    # set the maximum time (in seconds) we wait for workers and other processes to die during reload/shutdown
plugin = python3
max-requests = 1000 # Reload workers after the specified amount of managed requests (avoid memory leaks).
vacuum = True       # Try to remove all of the generated files/sockets (UNIX sockets and pidfiles) upon exit.
module = kolibri.deployment.default.alt_wsgi:alt_application
buffer-size = 8192  # needed to support long requests blocks (in lessons, for example)
wsgi-disable-file-wrapper = true  # needed to fix uwsgi bug https://github.com/unbit/uwsgi/issues/1126
single-interpreter = true
die-on-term = true                   # Shutdown when receiving SIGTERM (default is respawn)
need-app = true

disable-logging = true               # Disable built-in logging
log-4xx = true                       # but log 4xx's anyway
log-5xx = true                       # and 5xx's

py-callos-afterfork = true           # allow workers to trap signals

max-requests = 1000                  # Restart workers after this many requests
max-worker-lifetime = 3600           # Restart workers after this many seconds
worker-reload-mercy = 60             # How long to wait before forcefully killing workers

# algorithm to scale automatically:
cheaper-algo = spare # set cheaper algorithm to use, if not set default will be used
cheaper = 2          # minimum number of workers to keep at all times
cheaper-initial = 3  # number of workers to spawn at startup
workers = 16         # maximum number of workers that can be spawned
cheaper-step = 1     # how many workers should be spawned at a time

manage-script-name = true