Merge branch '165-hardware-requirements' into 'develop'

Resolve "Document hardware requirements (CPU/memory)"

Closes #165

See merge request funkwhale/funkwhale!141

deploy/docker-compose.yml

@@ -20,6 +20,14 @@ services:
     restart: unless-stopped
     image: funkwhale/funkwhale:${FUNKWHALE_VERSION:-latest}
     env_file: .env
+    # Celery workers handle background tasks (such file imports or federation
+    # messaging). The more processes a worker gets, the more tasks
+    # can be processed in parallel. However, more processes also means
+    # a bigger memory footprint.
+    # By default, a worker will span a number of process equal to your number
+    # of CPUs. You can adjust this, by explicitly setting the --concurrency
+    # flag:
+    #   celery -A funkwhale_api.taskapp worker -l INFO --concurrency=4
     command: celery -A funkwhale_api.taskapp worker -l INFO
     links:
       - postgres

deploy/funkwhale-worker.service

@@ -8,6 +8,14 @@ User=funkwhale
 # adapt this depending on the path of your funkwhale installation
 WorkingDirectory=/srv/funkwhale/api
 EnvironmentFile=/srv/funkwhale/config/.env
+# Celery workers handle background tasks (such file imports or federation
+# messaging). The more processes a worker gets, the more tasks
+# can be processed in parallel. However, more processes also means
+# a bigger memory footprint.
+# By default, a worker will span a number of process equal to your number
+# of CPUs. You can adjust this, by explicitly setting the --concurrency
+# flag:
+#   celery -A funkwhale_api.taskapp worker -l INFO --concurrency=4
 ExecStart=/srv/funkwhale/virtualenv/bin/celery -A funkwhale_api.taskapp worker -l INFO
 
 [Install]

docs/installation/index.rst

@@ -12,9 +12,48 @@ The project relies on the following components and services to work:
 - A celery worker to run asynchronouse tasks (such as music import)
 - A celery scheduler to run recurrent tasks
 
+
+Hardware requirements
+---------------------
+
+Funkwhale is not especially CPU hungry, unless you're relying heavily
+on the transcoding feature (which is basic and unoptimized at the moment).
+
+On a dockerized instance with 2 CPUs and a few active users, the memory footprint is around ~500Mb::
+
+   CONTAINER                   MEM USAGE
+   funkwhale_api_1             202.1 MiB
+   funkwhale_celerybeat_1      96.52 MiB
+   funkwhale_celeryworker_1    168.7 MiB
+   funkwhale_postgres_1        22.73 MiB
+   funkwhale_redis_1           1.496 MiB
+
+Thus, Funkwhale should run fine on commodity hardware, small hosting boxes and
+Raspberry Pi. We lack real-world exemples of such deployments, so don't hesitate
+do give us your feedback (either positive or negative).
+
+Software requirements
+---------------------
+
+Software requirements will vary depending of your installation method. For
+Docker-based installations, the only requirement will be an Nginx reverse-proxy
+that will expose your instance to the outside world.
+
+If you plan to install your Funkwhale instance without Docker, most of the
+dependencies should be available in your distribution's repositories.
+
+.. note::
+
+   Funkwhale works only with Pyhon >= 3.5, as we need support for async/await.
+   Older versions of Python are not supported.
+
+
 Available installation methods
 -------------------------------
 
+Docker is the recommended and easiest way to setup your Funkwhale instance.
+We also maintain an installation guide for Debian 9.
+
 .. toctree::
    :maxdepth: 1