[metrics-es-node-graphite] jvm_stats max_heap_size_in_bytes already done by Elasticsearch

[chrisduong]

Hi,

I'm using Elasticsearch v1.4.4 and I like this plugin. I recognized that the "jvm_stats max_heap_size_in_bytes" has been already done by ES. The "heap_max_in_bytes" has the same value.

node['jvm']['mem']['heap_max_in_bytes']

And one more, adding "used_in_bytes" for each pool is also useful, so we can see how GC work in real time, we can see the "GC frequency and duration" as from ES official guide . I had patched it for myself.

      node['jvm']['mem']['pools'].each do |k, v|
        # get `used_in_bytes`.
        metrics["jvm.mem.#{k.tr(' ', '_')}.used_in_bytes"] = v['used_in_bytes']
        metrics["jvm.mem.#{k.tr(' ', '_')}.max_in_bytes"] = v['max_in_bytes']

[majormoses]

@chrisduong sounds like something we want, do you have a pr for review?