[skip ci] Making more clear what to configure

Author: Chris Wiechmann

README.md

@@ -720,10 +720,10 @@ The following table explains what a single component, such as Logstash, Filebeat
 | Component               | Max. TPS           | Host-Machine | Config                | Comment |
 | :---                    | :---               | :---         | :---                  | :---    | 
 | Filebeat                | >300               | t2.xlarge    | Standard              | Test was limited by the TPS the Mock-Service was able to handle. Filebeat can very likely handle much more volume.|
-| Logstash                | 530                | t2.xlarge    | 6GB RAM for Logstash  | Includes API-Builder & Memcache on the same machine running along with Logstash. Has processed ap. 3500 events per second. CPU is finally the limiting factor.  A production setup should have two Logstash nodes for high availability, which provides sufficient capacity for most requirements.|
-| 2 Elasticsearch nodes   | 480                | t2.xlarge    | 8GB RAM for each node | Starting with a Two-Node cluster as this should be the mininum for a production setup. Kibana running on the first node.|
-| 3 Elasticsearch nodes   | 740                | t2.xlarge    | 8GB RAM for each node | Data is searchable with a slight delay, but ingesting is not falling behind real-time in general up to the max. TPS.|
-| 4 Elasticsearch nodes   | 1010                | t2.xlarge    | 8GB RAM for each node | |
+| Logstash                | 530                | t2.xlarge    | 6GB JVM-Head for Logstash  | Includes API-Builder & Memcache on the same machine running along with Logstash. Has processed ap. 3500 events per second. CPU is finally the limiting factor.  A production setup should have two Logstash nodes for high availability, which provides sufficient capacity for most requirements.|
+| 2 Elasticsearch nodes   | 480                | t2.xlarge    | 8GB JVM-Heap for each node | Starting with a Two-Node cluster as this should be the mininum for a production setup. Kibana running on the first node.|
+| 3 Elasticsearch nodes   | 740                | t2.xlarge    | 8GB JVM-Heap for each node | Data is searchable with a slight delay, but ingesting is not falling behind real-time in general up to the max. TPS.|
+| 4 Elasticsearch nodes   | 1010                | t2.xlarge    | 8GB JVM-Heap for each node | |
 
 Please note:  
 - Logstash, API-Builder, Filebeat (for monitoring only) and Kibana are load balanced across all available Elasticsearch nodes. An external Load-Balancer is not required as this is handled internally by each of the Elasticsearch clients.