CPU

field name: cpu_some — what munin.conf and the RRD call it
unit: % of time stalled
state: OK
thresholds: no warning · no critical
limits: ≥ 0 — values outside this range are stored as UNKNOWN rather than clipped (rrdcreate's DS min/max)
about: Share of time at least one task was stalled waiting for CPU.
drawn as: cpu_some,10000,/ — the figures below plot this, not the stored value
data source: localhost/localhost-psi-cpu_some-d.rrd · DERIVE
graphs: Pressure stall · zoom

24-hour profile (daily model, mean ± σ, server-local time — CEST, UTC+02:00) · the second line is the last 24 hours, ending at the marker

the numbers behind this figure — 48 half-hour slots

unit: % of time stalled · typical is the daily model's mean at that local half-hour, band its mean ∓ σ held inside the field's declared range, n the samples pooled behind it, and last 24 h what the field actually did on the same slot
slot	typical	band	n	last 24 h
00:00	3.81	1.13 … 6.49	15	3.19
00:30	3.76	1.06 … 6.46	15	3.47
01:00	2.62	1.97 … 3.27	15	2.38
01:30	2.21	2.09 … 2.34	15	2.24
02:00	2.17	2.09 … 2.26	15	2.31
02:30	2.29	2.02 … 2.56	15	2.13
03:00	3.21	1.87 … 4.56	15	2.75
03:30	3.20	1.84 … 4.55	15	4.24
04:00	3.04	1.60 … 4.49	15	2.11
04:30	2.05	1.88 … 2.22	15	1.91
05:00	1.99	1.78 … 2.20	15	1.51
05:30	2.10	1.58 … 2.62	15	1.53
06:00	2.59	823.55m … 4.35	15	3.83
06:30	2.60	834.27m … 4.36	15	8.72
07:00	2.48	746.32m … 4.22	15	1.43
07:30	2.02	1.84 … 2.21	15	2.02
08:00	2.05	1.98 … 2.13	15	2.30
08:30	2.07	1.98 … 2.15	15	2.04
09:00	2.07	2.00 … 2.14	15	2.05
09:30	2.08	2.00 … 2.15	15	2.05
10:00	2.11	1.94 … 2.27	15	1.99
10:30	2.11	1.95 … 2.28	15	2.65
11:00	2.14	1.94 … 2.33	15	2.06
11:30	2.11	1.96 … 2.25	15	1.98
12:00	2.08	1.94 … 2.23	15	2.02
12:30	2.08	1.93 … 2.23	15	1.98
13:00	2.12	1.92 … 2.32	15	2.58
13:30	2.15	1.96 … 2.34	15	2.04
14:00	2.12	1.97 … 2.27	15	2.14
14:30	2.15	1.94 … 2.37	15	2.07
15:00	2.26	1.79 … 2.72	15	2.14
15:30	2.47	1.59 … 3.36	15	3.71
16:00	2.58	1.51 … 3.64	15	5.24
16:30	2.58	1.52 … 3.64	15	4.68
17:00	2.45	1.67 … 3.24	15	3.69
17:30	2.35	1.84 … 2.87	15	3.18
18:00	2.31	1.87 … 2.75	15	3.00
18:30	2.33	1.88 … 2.77	15	3.21
19:00	2.59	1.51 … 3.66	16	3.26
19:30	2.66	1.58 … 3.74	17	3.27
20:00	2.59	1.54 … 3.64	18	3.39
20:30	2.39	1.80 … 2.98	17	2.72
21:00	2.34	1.84 … 2.84	16	3.63
21:30	2.42	1.81 … 3.04	15	3.39
22:00	2.45	1.84 … 3.06	15	3.66
22:30	2.38	1.82 … 2.95	14	3.56
23:00	2.25	1.92 … 2.58	15	2.46
23:30	3.38	578.33m … 6.19	15	2.09

hour × weekday (weekly model mean, server-local time — CEST, UTC+02:00)

the numbers behind this figure — 7 weekdays × 12 buckets

unit: % of time stalled · each cell is the weekly model's mean over that local bucket, with n the samples pooled behind it — the same value and pool the coloured cell above carries
day	00:00	02:00	04:00	06:00	08:00	10:00	12:00	14:00	16:00	18:00	20:00	22:00
Mon	2.45 n=12	2.67 n=12	2.51 n=12	2.39 n=12	2.18 n=12	2.17 n=12	2.27 n=12	2.28 n=12	2.34 n=12	2.33 n=12	2.30 n=12	2.42 n=12
Tue	2.61 n=12	2.85 n=12	2.69 n=12	2.39 n=12	2.11 n=12	2.13 n=12	2.18 n=12	2.23 n=12	2.28 n=12	2.29 n=12	2.32 n=11	2.41 n=10
Wed	2.54 n=9	2.75 n=8	2.65 n=7	2.51 n=6	2.21 n=6	2.20 n=6	2.57 n=6	2.58 n=6	2.58 n=7	2.36 n=8	2.32 n=9	2.36 n=9
Thu	2.42 n=9	2.74 n=9	2.66 n=9	2.51 n=9	2.22 n=9	2.23 n=9	2.20 n=9	2.33 n=9	2.38 n=9	2.68 n=10	2.57 n=11	2.65 n=12
Fri	2.53 n=12	2.79 n=12	2.63 n=12	2.50 n=12	2.20 n=12	2.20 n=12	2.20 n=12	2.23 n=12	2.26 n=12	2.20 n=12	2.15 n=12	2.62 n=12
Sat	2.81 n=12	3.12 n=12	2.58 n=12	2.37 n=12	2.07 n=12	2.10 n=12	2.16 n=12	2.21 n=12	2.19 n=12	2.18 n=12	2.14 n=12	2.24 n=12
Sun	2.39 n=12	2.86 n=12	2.88 n=12	2.95 n=12	2.47 n=12	2.36 n=12	2.17 n=12	2.49 n=12	2.69 n=12	2.74 n=12	2.60 n=12	2.50 n=11