Every day at the same hours we get a CPU spike that lasts for less then a minute.
Our request rate remains the same, watching marvel shows no increase at merge rate.
The main problem is that my latency grows and my backend starts degrade.
I'm using ES 1.7 and I have marvel installed. That's the only program that runs at my cluster machines.
I am thinking about some ES process or Garbage Collector or even something of Lucene...
I could enable jmx to watch Garbage Collector but I will need to reboot my cluster and that's not an option right now...
Do you have any idea of what is going on?
If you need more info let me know.