Fixes #509 - Updated hardware requirements

prerequisites/hardware.rst

@@ -1,41 +1,59 @@
 Hardware
 ********
 
-You can run Zammad on bare metal or on a virtual machine. Choose what 
-you prefer.
+The hardware requirements to run Zammad effectively vary depending on several
+factors, for example:
 
-For Zammad and a database server like PostgreSQL we recommend at least:
-=======================================================================
+- Number of users (especially agents and customers)
+- Volume of incoming tickets
+- Usage of channels
+- Installation type
+- Used software which is required to run Zammad
 
-* 2 CPU cores
-* 4 GB of RAM (+4 GB if you want to run Elasticsearch on the same server)
-
-For optimal performance up to 40 agents:
-========================================
+There are even many more factors which have influence on the performance. This
+makes it hard to publish general hardware requirements which are fitting
+for all use cases. If you don't want to deal with such things, you should
+consider our `hosted Zammad setup <https://zammad.com/en/pricing>`_.
 
-* 4 CPU cores
-* 6 GB of RAM (+6 GB if you want to run Elasticsearch on the same server)
+However, you can use the following information as a rough guide.
 
-Of course at the end it depends on actual load of concurrent agents and 
-data traffic.
+.. tip:: In any case,
+   more and better hardware ensures that Zammad runs smoothly and the
+   maintenance breaks for updates should be shorter as well.
 
-.. note:: 
+.. note::
 
-   We can't suggest any disk space recommendations, as this highly depends on 
-   how you work. Zammad will always try to recognize the same attachments and 
+   We can't make any disk space recommendations, as this highly depends on
+   how you work. Zammad will always try to recognize the same attachments and
    store it just once.
 
-Performance Tuning
-==================
+Minimum Setup
+=============
+
+For a minimalistic Zammad deployment with a PostgreSQL database server, you
+should not go below the following values:
+
+* 2 CPU cores
+* 6 GB of RAM (+4 GB if you want to run Elasticsearch on the same server)
+
+If you run your Elasticsearch instance on a lightweight machine like this and
+your RAM got exhausted, you could try a less-than-ideal solution like
+`limit the ES heap size <https://www.elastic.co/guide/en/elasticsearch/reference/current/advanced-configuration.html#set-jvm-heap-size>`_.
+
+Example Setup
+=============
+
+This is a basic setup for **up to 40 agents**. As mentioned above, it highly
+depends on many factors, but this can be a good starting point for dimensioning
+your system.
 
-As the number of active users on your system grows,
-performance will eventually degrade, leading to:
+* 4 CPU cores
+* 6 GB of RAM (+6 GB if you want to run Elasticsearch on the same server)
 
-   * delays for outgoing email,
-   * long loading times when viewing or creating tickets,
-   * stale or out-of-sync search results, or
-   * stale or out-of-sync ticket overviews.
+Performance Tuning
+==================
 
-You may see modest improvements by
-:ref:`setting certain environment variables for Performance Tuning <performance_tuning>`,
-such as ``$WEB_CONCURRENCY`` or ``$ZAMMAD_SESSION_JOBS_CONCURRENT``.
+As the number of active users and/or tickets grows, performance will
+eventually degrade. You should have a look at
+:ref:`performance tuning section <performance_tuning>` then, or consider upgrade
+your setup.