all, hello!I am a newcomer in ganglia.After i read the paper 'The ganglia distributed monitoring system',i have some questions.So i come here for help.Thanks a lot.Maybe someone can give me some clues which i use to find the right answers. 1.Every built-in metric has correspondent time thresholds,for example '900-1200' for cpu_num.But gmond expires monitoring data using timeouts.For each monitoring metric,it uses two time limits,a soft limit and a hard limit.What i am wondering is that what 's the relation between the time threshold,soft limit,hard limit.Are they equal? 2.what relation exists between node running gmetad and node running gmond.Can these two daemons must running on different nodes or they can exist in the same node? 3.the node running gmetad can get monitoring data from data sources. Data sources may be either gmond daemons, representing specific clusters, or other gmetad daemons, representing sets of clusters.does the communication between gmetad and gmetad differ from the communication between gmetad and gmond.Using the same protocol (XML over TCP) or the same interface? 4.Does ganglia has the same event mechinism like the trap in SNMP? 5.Each built-in metric has collect/announce schedule, metric value thresholds, metric units and binary to text conversion details. These values can be modified at compile time to accommodate different environments.Can these values be modified when running?For example,I can use some monitored data to change these value by some policies. 6.If i wanna read the source code of ganglia,where can i start?Can anyone give me some clues or instructions.
Thanks a lot lot again!
