bottleneck prints out load averages for CPU, memory, disk and network. These are saturation levels only. In this example we run bottleneck with 5 second samples. As bottleneck runs, load is applied to the server in different areas. It was necessary to saturate the resources for load to register, these values do not indicate utilisation - rather saturation. At 02:28:35, several CPU intensive applications were run and then stopped a short time later. At 02:29:20, an application is executed that used a lot of of memory (also causing disk load). At 02:30:08, several programs were run that made heavy use of the disks. And finally at 02:30:53, a few programs were executed that generated heavy network traffic. All of this activity can be seen in the load averages. $ ./bottleneck 5 Time CPU Mem Disk Net 02:28:27 0.00 0.00 0.00 0.00 02:28:32 0.00 0.00 0.01 0.00 02:28:37 0.80 0.00 0.00 0.00 02:28:42 3.80 0.00 0.00 0.00 02:28:47 4.00 0.00 0.00 0.00 02:28:52 4.00 0.00 0.00 0.00 02:28:57 2.60 0.00 0.00 0.00 02:29:02 0.00 0.00 0.00 0.00 02:29:07 0.00 0.00 0.00 0.00 02:29:12 0.00 0.00 0.00 0.00 02:29:17 0.40 0.00 0.00 0.00 02:29:26 2.00 17.00 2.73 0.00 02:29:31 0.12 9.86 4.18 0.00 02:29:36 0.00 0.43 0.22 0.00 02:29:41 0.00 0.00 0.00 0.00 02:29:46 0.00 0.00 0.01 0.00 02:29:51 0.00 0.00 0.00 0.00 02:29:56 0.60 0.00 0.17 0.00 02:30:01 2.00 0.00 0.00 0.00 02:30:06 2.80 0.00 0.00 0.00 Time CPU Mem Disk Net 02:30:11 1.00 0.00 0.69 0.00 02:30:16 1.20 0.00 1.16 0.00 02:30:21 0.40 0.00 1.72 0.00 02:30:26 0.00 0.00 2.46 0.00 02:30:31 0.00 0.00 1.86 0.00 02:30:36 1.20 0.00 1.88 0.00 02:30:41 2.40 0.00 1.85 0.00 02:30:46 1.20 0.00 0.51 0.00 02:30:51 0.00 0.00 0.00 0.00 02:30:56 0.00 0.00 0.00 1.91 02:31:01 0.60 0.00 0.00 3.91 02:31:06 0.40 0.00 0.00 2.73 02:31:11 0.00 0.00 0.00 0.00 02:31:16 0.00 0.00 0.00 0.00 02:31:21 0.00 0.00 0.00 0.00 ^C bottleneck may be useful when left running in a window to always provide the 1 minute averages, $ ./bottleneck 60 Time CPU Mem Disk Net 12:46:51 0.02 0.00 0.00 0.00 12:47:51 0.02 0.00 0.00 0.00 12:48:51 0.00 0.00 0.00 0.00 12:49:51 0.02 0.00 0.00 0.00 12:50:51 0.00 0.00 0.00 0.00 [...] The above indicates a quiet system. Ever noticed how installing the gcc-3.3 compiler hammers the disks? Out of curiosity I ran bottleneck as I installed gcc on a system, over an NFS share. We can clearly see disk saturation, and even a little CPU saturation at the start as the gcc package was decompressed and processed by pkgadd (pkgadd was reading the output of gzcat), $ ./bottleneck 10 Time CPU Mem Disk Net 18:22:00 0.00 0.00 0.00 0.00 18:22:10 0.00 0.00 0.00 0.00 18:22:20 0.20 0.00 1.05 0.00 18:22:30 0.50 0.00 0.02 0.00 18:22:40 1.00 0.00 0.00 0.00 18:22:50 0.60 0.00 4.07 0.00 18:23:00 0.30 0.00 0.00 0.00 18:23:10 0.80 0.00 58.84 0.00 18:23:20 0.00 0.00 1.51 0.00 18:23:30 0.00 0.00 0.77 0.00 18:23:40 0.00 0.00 0.99 0.00 18:23:50 0.00 0.00 3.10 0.00 18:24:00 0.00 0.00 4.96 0.00 18:24:10 0.00 0.00 2.64 0.00 18:24:20 0.00 0.00 62.29 0.00 18:24:30 0.00 0.00 5.91 0.00 18:24:40 0.00 0.00 0.79 0.00 18:24:50 0.00 0.00 1.35 0.00 18:25:00 0.00 0.00 1.23 0.00 18:25:10 0.00 0.00 2.88 0.00 Time CPU Mem Disk Net 18:25:20 0.00 0.00 4.07 0.00 18:25:30 0.00 0.00 0.33 0.00 18:25:40 0.00 0.00 0.00 0.00 18:25:50 0.00 0.00 2.84 0.00 18:26:00 0.00 0.00 0.00 0.00 18:26:10 0.00 0.00 0.00 0.00 18:26:20 0.00 0.00 0.00 0.00 For more examples of these load averages, see the individual load programs and their examples: loadcpu, loadmem, loaddisk and loadnet.