Hi Werner, each Graylog node (using the Graylog Graphite/Ganglia output plugin <https://github.com/Graylog2/graylog-plugin-metrics>) will only send the messages to Graphite, which it received and processed.
So if you're using a round-robin mechanism to distribute messages to the Graylog nodes, each Graylog node will process and output a distinct subset of messages. This subset of messages is being batched for 60 seconds (in your case) and then sent to Graphite as distinct data points. Cheers, Jochen On Monday, 21 November 2016 03:44:51 UTC+1, Werner van der Merwe wrote: > > Hi, > > We are running 3 Graylog nodes behind a round-robin load balancer that > balancer all comms to Graylog over the three nodes. > > I was curious how the metrics output plugin manages this. The output > plugin is set to send data every 60 seconds. Graphite storage schema is set > to 1 minute as well. Thus if all three nodes send the data, they'd be > averaged by Graphite rather than combined, resulting in incorrect metrics > being plotted. > > I am not able to confirm or deny the behaviour, but I would like to know > more than 'best-guess' on my part. > > Thanks! > -- You received this message because you are subscribed to the Google Groups "Graylog Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to graylog2+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/graylog2/5051c570-5b38-4699-82f1-a642796ac151%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
