What does Pocket Command of Container Fleets conclude, in numbers, about “Grouping and Interruption: One Event, One Push”?
Answer: Chapter 3's three-stage grouping cuts pushes from 214.2 a day to 1.30, a reduction of 0.9939.
Scope of the numbers: Its parameters are illustrative; all calculations are generated by its programs; the book offers non-personalised operating guidance, promises no availability figure, and was written without connecting to any machine. Marked boxes separate published evidence, runnable Python calculations (22 programs), classical sources and the author's hypotheses.
Data table
The numbers below are printed verbatim by the chapter's Python program and can be reproduced by running it. standard-library Python onlyprinted values that match the answer: 0.9939, 1.30, 214.2
| real faults in 30 days | 35 |
| raw events | 6,425 |
| after same-source dedup | 1,326 |
| after 10-min grouping | 744 |
| pushed (corroborated) | 39 |
| sent to digest | 705 |
| real-fault groups digested | 0 |
| faults covered by a push | 35 |
| pushes per day 1.30; raw per day 214.2 | |
| reduction raw -> pushed: 0.9939 |
Show the full program
# Chapter 3. p31: same-source dedup and time-window grouping of alerts
# Thirty synthetic days (fixed seed). Real faults arrive about 1.2 a
# day; each hits one container and fans out to 0 to 8 dependants, and
# every affected (container, check) pair re-fires each minute while the
# fault lasts. Isolated one-shot flaps add about 30 events a day.
# Stage 1: drop a repeat of the same (container, check) within 30 min.
# Stage 2: join events less than 10 min after the open group's last one.
# Stage 3: push a group only when some container in it fails two or
# more distinct checks; other groups go to the digest.
import math
import random
rng = random.Random(41631)
MIN = 30 * 24 * 60
events, faults = [], []
t = 0.0
while True:
t += rng.expovariate(1.2 / 1440.0)
if t >= MIN:
break
root = rng.randrange(40)
dur = max(2, int(math.exp(rng.gauss(math.log(12.0), 0.7))))
hit = [root] + [rng.randrange(40) for _ in range(rng.randint(0, 8))]
faults.append(int(t))
for c in sorted(set(hit)):
for chk in range(rng.randint(2, 4)):
for m in range(dur):
events.append((int(t) + m, c, chk, 1))
for _ in range(30 * 30):
events.append((rng.randrange(MIN), rng.randrange(40), 9, 0))
events.sort()
last, stage1 = {}, []
for ev in events:
key = (ev[1], ev[2])
if key in last and ev[0] - last[key] < 30:
continue
last[key] = ev[0]
stage1.append(ev)
groups = []
for ev in stage1:
if groups and ev[0] - groups[-1][-1][0] < 10:
groups[-1].append(ev)
else:
groups.append([ev])
def corroborated(g):
seen = {}
for ev in g:
seen.setdefault(ev[1], set()).add(ev[2])
return any(len(v) >= 2 for v in seen.values())
pushed = [g for g in groups if corroborated(g)]
digest = [g for g in groups if not corroborated(g)]
lost = [g for g in digest if any(ev[3] == 1 for ev in g)]
covered = 0
for f in faults:
if any(g[0][0] <= f <= g[-1][0] + 1 for g in pushed):
covered += 1
print("real faults in 30 days {:>8,}".format(len(faults)))
print("raw events {:>8,}".format(len(events)))
print("after same-source dedup {:>8,}".format(len(stage1)))
print("after 10-min grouping {:>8,}".format(len(groups)))
print("pushed (corroborated) {:>8,}".format(len(pushed)))
print("sent to digest {:>8,}".format(len(digest)))
print("real-fault groups digested {:>8,}".format(len(lost)))
print("faults covered by a push {:>8,}".format(covered))
print("pushes per day {:.2f}; raw per day {:,.1f}".format(
len(pushed) / 30.0, len(events) / 30.0))
cut = 1.0 - len(pushed) / len(events)
print("reduction raw -> pushed: {:.4f}".format(cut))Single source and citation
Fan, Y. (2026). Pocket Command of Multi-Container Fleets: Mobile Remote Control, Exception-Only Push Filtering and Away-From-Keyboard Continuity (Version 1.0.0), Chapter 3: Grouping and Interruption: One Event, One Push. 5D Library. DOI pending
DOI pending · monograph page goes live after publication · 5DL-B-416 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026416,
author = {Fan, Yafu},
title = {Pocket Command of Multi-Container Fleets: Mobile Remote Control, Exception-Only Push Filtering and Away-From-Keyboard Continuity},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}Other chapter answers in this book
- Chapter 1 · Base Rates: the More Pushes, the Less Credible Each One
- Chapter 2 · Detectors: Cumulative Sums, Weighted Averages and the Error Budget
- Chapter 4 · Heartbeats and Probes: Late Is Not Dead
- Chapter 5 · Automatic Repair: Backoff, Circuit Breakers and the Availability Ledger
- Chapter 6 · Lost Links, Thin Links and Permissions: What the Phone May and May Not Do
- Chapter 7 · The Tent, the Field and Requisite Variety: Eastern and Western Sources Examined
- Chapter 8 · Field Tools and Checker Specifications: the Away-From-Desk Ledger and a Pre-Registered Test