Logstash throughput with metrics plugin

Hi everybody,

I am trying to measure the logstash's throughput with the metrics plugin and I don't understand the results I obtained, can anyone help me?

My configuration:

forwarders (~6000 events/min) -> Logstash/REDIS(Node02) -> Logstash/Indexer(Node04) -> ElasticSearch(Node06)

In the Logstash node that adds metrics to Redis I added a metrics output to see the input rate. In the indexer I did the same to see the output rate of the agent. I'm sending
the metrics to graphite and I am observing that although when the input rate it's the same that output rate, the redis queue grows. When the queue Grows, the output rate is limited to 50 events/sec in the indexer.
Could anyone explain what is happening or what I am doing wrong?

Thanks in advance