Hi all,
This is slightly tricky - well for me anyways..
I have an index where a key event is occuring. I need to be able to examine the 5 events prior to and after the key event (irrespective of time) per host, and tabulate a count of each event type (by type i mean the string that appears in the log entry).. so if we have 7 distinct event types surrounding all instances of the key event, then we would see a table with 7 rows, showing the event and a count of occurances.
The idea is to use proximal analysis to get a bead on possible causal relationship. I'm pretty sure this is a common practice - but after hunting around I can't find a reference to it anywhere..
Would much appreciate any Ninja feedback..
Thx.