The "ANY" percentages refer to combinations of CPU cores - ANY1+ is time that any single core is busy, ANY2+ is time that any two cores are busy, etc. The two values - single CPU % busy and the ANYs give you a better picture of the overall system load.
For instance, if each individual CPU core indicated 75% busy during the sample interval, but the ANY3+ or so was only 10%, then the cores aren't really busy at the same time. Either the parallelism is spreading the load around well, or there are serialized loads limiting the ability to utilize maximum compute capability. The CPU values give you directions in which to look for specific performance issues.
Similarly, of the individuals CPU cores are around 75% and in your example say ANY5+ was pushing 70-75%, then you clearly have a system approaching full CPU load - good spread in parallel task execution but obviously a lot of them at one time.
The rest of the columns are specific processing domains in which stuff happens. Kahuna and WAFL_EX(Kahu) are related in that they are mutually exclusive to each other. Kahuna tasks are single threaded, WAFL_EX tasks can be run in parallel.