1loads(8) System Manager's Manual loads(8)
2
3
4
6 loads.bt - Prints load averages. Uses bpftrace/eBPF.
7
9 loads.bt
10
12 These are the same load averages printed by "uptime", but to three dec‐
13 imal places instead of two (not that it really matters). This is really
14 a demonstration of fetching and processing a kernel structure from bpf‐
15 trace.
16
17 Since this uses BPF, only the root user can use this tool.
18
20 CONFIG_BPF and bpftrace.
21
23 Print system load averages every second:
24 # loads.bt
25
27 HH:MM:SS
28 Each output line includes time of printing in "HH:MM:SS" format.
29
30 load averages:
31 These are exponentially-damped moving sum averages of the system
32 loads. Load is a measurement of demand on system resources,
33 which include CPUs and other resources that are accessed with
34 the kernel in an uninterruptible state (TASK_UNINTERRUPTIBLE),
35 which includes types of disk I/O and lock accesses. Linux load
36 averages originally reflected CPU demand only, as it does in
37 other OSes, but this was changed in Linux 0.99.14. This demand
38 measurement reflects not just the utilized resource, but also
39 the queued demand (a saturation measurement). Finally, the three
40 numbers are called the "one-", "five-", and "fifteen-minute"
41 load averages, however these times are constants used in the ex‐
42 ponentially-damping equation, and the load averages reflect load
43 beyond these times. Were you expecting an accurate description
44 of load averages in the man page of a bpftrace tool?
45
47 Other than bpftrace startup time, negligible.
48
50 This is from bpftrace.
51
52 https://github.com/iovisor/bpftrace
53
54 Also look in the bpftrace distribution for a companion _examples.txt
55 file containing example usage, output, and commentary for this tool.
56
58 For more on load averages, see:
59
60 http://www.brendangregg.com/blog/2017-08-08/linux-load-averages.html
61
63 Linux
64
66 Unstable - in development.
67
69 Brendan Gregg
70
72 uptime(1)
73
74
75
76USER COMMANDS 2018-09-10 loads(8)