How to know which script or executable is linked with a metric in ganglia?

Question

I have just started to explore ganglia and my question is "How to know which script or executable is linked with a metric in ganglia?"

The fact is that I don't know much about ganglia. I have good experience in zabbix and I want to link a graph in zabbix which is already working in ganglia.

I have browsed a lot on the Internet but none of the answers was relevant.

Suppose, the metric is:

"hbase.RegionServerDynamicStatistics.tbl.SUBSCRIBER.cf.M_T_V"

Beneath this, 2 graphs are shown as "hbase.RegionServerDynamicStatistics.tbl.SUBSCRIBER.cf.M_T_V.get__avg_time" and "hbase.RegionServerDynamicStatistics.tbl.SUBSCRIBER.cf.M_T_V.get__num_ops"

As far as I understand, this metric is not inbuilt. Correct me if I am wrong.

There must be some script or executable collecting these values. So, please help me to find the script?

I have also tried to find such metrics in the configuration files, such as, /etc/ganglia/gmond.conf and /etc/ganglia/gmetad.conf files, but not any such entry was there.

I read somewhere that some custom metric command line could be formed based on gmetric. And that could be attached in cron job, which would send data to the ganglia server periodically. If this is the way it happens then is there any way to find the gmetric based command line formed to do so?

Answer 1

Ok, I got the answer of this question. This could be helpful for others too.

Actually, hadoop has a class named as GangliaContext31.java which emits all hadoop related metrics to the ganglia server based on the IP or FQDN of the ganglia server, provided in the configuration file of hadoop known as hadoop-metrics2.properties.

A short excerpt of this configuration file is shown as:

namenode.sink.ganglia.class=org.apache.hadoop.metrics2.sink.ganglia.GangliaSink31
namenode.sink.ganglia.period=10
namenode.sink.ganglia.servers=ganglia-server:8649 

Here ganglia-server is the said FQDN.

gmond daemon installed on ganglia-server collects these metrics and forward them to gmetad daemon installed on the ganlia-server. gmetad daemon stores these metrics in the rrdtool . In this way, graphs are built based on these values stored in the rrdtool.