On Tue, 2013-01-15 at 08:28 -0800, Craig Cook wrote: > > Graphite/Carbon - missing data correlation > > > What do you mean by "missing data correlation"? > > > Something like, "show me CPU usage across my web cluster?" > > > Or, "See that CPU spike on Server A, is it related to the slow > response of App B?" > Robert Au mentioned in his post, opentsdb, which I think is a good candidate.
The questions usually hover around maximum and average latency in different components. With correlation I mean the ability to have data from different "tables" or "Databases" considered to answer a question. For example, Q1: what was the maximum number of interrupts/second on CPU #15 during the past 24h. Q2: what was the maximum number of interrupts/second on Server X (which has CPU 0 to 63) during the past 24h. Q3: what was the maximum number of interrupts/second on any Server in Datacenter X during the past 24h. And so on.... This kind of data I want for network/storage/cpu/ram usage and what have you not. It is realistic to assume there will be correlation of different types, e.g. how do IRQ/s correlate with IOPs (if at all ;) ) I hope your question is answered, and again thanks everyone for the valuable suggestions! Conrad _______________________________________________ Tech mailing list [email protected] https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech This list provided by the League of Professional System Administrators http://lopsa.org/
