The Quiet Month Report: What to Show When Nothing Broke
The month after an outage, the report writes itself. There is a timeline, a cause, a fix, and a list of things you are doing so it never happens again. Everyone reads it because everyone remembers the afternoon. The month when nothing broke is the opposite. The page stays empty, because the honest summary is "it was fine", and "it was fine" is not a report anyone forwards.
The quiet month is the hardest report to write and the most important one, because it is the only place the quiet work ever appears. An outage report proves you can recover. A quiet month report proves you can prevent, and prevention is the thing you are actually paid for. I wrote "no incidents this month" for two years before someone gently asked what I did with the rest of the time.
By the end of this guide you will be able to write the quiet month report in four fixed sections. You will be able to pull each section from your monitoring system rather than from memory. You will know how to leave out the things that make it read like busywork and keep the tone plain enough to be trusted. And you will be able to produce the page in under an hour. There is a template to copy and a filled example to pattern on.
In this guide
- Why the quiet month is the hardest report
- The four sections
- What to leave out
- Tone rules
- How it fits with the monthly one-pager
- The template
- A filled example: Harbourview Clinic
- Producing it in under an hour
- The report that changed the question
- The checklist
- Where to go next
Why the quiet month is the hardest report
A quiet month has no story, and reports are usually built around a story. An outage gives you a beginning, a middle, and an end. A quiet month gives you thirty days in which forty small things were dealt with before they became stories. None of the forty can carry a page on its own. So the temptation is to write nothing, or to write "no incidents", which is the same thing with a heading.
The quiet month report solves this by replacing the story with a fixed shape. Four sections, the same four every month, each filled from a specific place in the monitoring system. The shape does the work the story would have done. The reader knows where to look, what each section means, and how this month compares to last. A fixed shape is also what makes the report cheap, because you are filling boxes rather than composing.
The invisible work problem explains why the quiet month looks idle and introduces the work log, a one-line-per-event record kept on the day. This guide assumes you have something like that log. If you do not, start it today and write this month's report from the alert history alone; it will be thinner, and next month's will not be.
The four sections
The report has four sections, in this order, each sourced from monitoring data rather than recollection. The order runs from the most impressive thing you did to the thing you need help with. So a reader who stops early leaves with the right impression, and a reader who reaches the end knows what you want.
What we caught early
This section lists the problems the monitoring system showed before they became outages, and what you did about each. The source is the work log and, if you keep one, the prevented-incident log. Each entry is one line: the signal, the action, and what would have happened. "File server volume at 94 per cent and rising; extended by 500 GB on the 12th; would have filled within the week and stopped the shared drive." Trend and capacity catches belong here; most come out of the weekly trend review, with the buying decisions timed by lead time. The rules for what counts as a genuine catch are in counting the outages you prevented. The short version is that an alert that cleared on its own never appears here.
What we handled without anyone noticing
This section is the alerts a person acknowledged and resolved during the month. It has two parts: the count, and the three entries worth a sentence each. The count comes from the alert history and sits beside last month's count. The three sentences are chosen for what they would have become, not for how hard they were. One might cover the mail queue restarted at six in the morning. Another might cover the flapping port fixed before the finance printer went offline. A third might cover the backup that failed on Tuesday night and was re-run before anyone needed it. The reader should finish this section knowing that things went wrong every week and nobody outside IT ever felt it.
What we changed safely
This section is the maintenance windows completed without incident. A maintenance window is a pre-announced period in which a change is made and alerts for the affected devices are suppressed. The source is the change record and the alert history in the hour after each window closed. State how many windows were planned, how many completed, and how many produced an alert afterwards. It is the section most often left out because nothing happened, which is precisely the point. Maintenance windows and suppression covers how windows are tracked, and the report can be pulled from there.
What is coming
This section is the trends that will need a decision, with a date. The source is the weekly trend review and the capacity register. "Internet link peak utilisation rising two per cent a month; upgrade decision needed by the end of next quarter." This is where the report earns its next reading, because it tells the manager what will be asked of them and when. It also makes next month's "caught early" section easier to write, because you will have said in advance what you were watching.
Beneath the four sections is one line: the availability of each critical service this month. That is the percentage of the time the service was working when it was supposed to be. Pull it from the availability report, exclude planned downtime, and state the method once. Honest availability math covers how to calculate a number that survives being checked; the report only has to quote it.
What to leave out
The quiet month report is short because most of what happened does not belong in it. Three things get left out on purpose.
- The raw alert count. "The monitoring system raised 1,140 alerts this month" is a statement about your thresholds, not your work. It invites the question "why so many?", which is fair and has a long answer. Report acknowledged and resolved, not raised.
- Anything that reads as busywork. Patches applied, tickets closed, passwords reset, printers unjammed. That work is real and visible elsewhere; here it dilutes the point. If a patch prevented something specific, it goes in "caught early" with the specific thing.
- Graphs, unless one graph is the story. One graph is allowed when a catch is best shown as a line climbing towards a limit and dropping on the day you acted. Two graphs is a slide deck. Reports people actually read is about why shorter wins, and everything it says applies here twice.
The test for any line is whether removing it would make the reader less sure that the team is looking after the infrastructure. If it would not, remove it. The report is not a record of effort; it is a record of things that did not happen.
Tone rules
The report is trusted in proportion to how plain it is.
- Plain words. "Fixed", "replaced", "extended", "renewed". Not "remediated", "actioned", or "leveraged". The reader should be able to read the report aloud without slowing down.
- Short entries. One line per item: signal, action, consequence avoided. If an item needs a paragraph, it belongs in a separate note with a link.
- No adjectives. Not "critical", "significant", "proactive", or "robust". "Would have stopped invoicing" is stronger than "critical" because the reader can picture it.
- One dry line at most. A single wry sentence in the covering email is fine. Inside the report, none. The report will be forwarded to someone who does not know you.
These rules feel restrictive the first month and invisible by the third. What they buy you is a report that reads the same whether the month was quiet or terrible. So the reader learns to trust the shape rather than the mood. A report that only sounds calm in calm months is not calm; it is weather.
How it fits with the monthly one-pager
The quiet month report and the monthly one-pager are two views of the same month for two readers. The monthly one-pager, with its four boxes, is the finance version. It goes to the person who signs the budget and is written in outcomes and risks. The quiet month report is the IT-team version: it goes to your manager and stays in the team's folder, written in signals and actions. The one-pager is built from the quiet month report, not the other way round.
If you only have time for one, write the quiet month report. It is the record; the one-pager is a summary of the record, and a summary of nothing is nothing. Nobody has ever complained about receiving a short report twice.
The template
Copy this into a text file and fill it in once a month. Keep every heading even when a section is empty; write "none this month" rather than deleting it.
QUIET MONTH REPORT - <team> - <month> 1. WHAT WE CAUGHT EARLY - <signal> on <service>: <action taken on date>; would have <consequence>. (source: work log, prevented-incident log) 2. WHAT WE HANDLED WITHOUT ANYONE NOTICING Alerts acknowledged and resolved by a person: <count> (last month: <count>) Three worth a sentence: - <date>: <what happened>; <what we did>; <who would have noticed>. (source: alert history) 3. WHAT WE CHANGED SAFELY Maintenance windows: <completed> of <planned>; follow-up alerts: <count> - <change>: <window date>; <result>. (source: change record; alert history after each window) 4. WHAT IS COMING - <service or resource>: <trend>; decision needed by <date>. (source: weekly trend review; capacity register) AVAILABILITY THIS MONTH (planned downtime excluded) <service>: <xx.xx%> <service>: <xx.xx%> <service>: <xx.xx%> Method: <one sentence> Next report: <date>
A filled example: Harbourview Clinic
Harbourview Clinic is a fictional outpatient clinic with one administrator. It has three critical services — patient records, the appointment booking site, and the internet link — and about ninety monitored devices. The numbers are illustrative; the shape is the point.
QUIET MONTH REPORT - Clinic IT - <month>
1. WHAT WE CAUGHT EARLY
- Records database volume at 91% and rising ~1%/day: extended by 200 GB
on the 9th; would have filled by the 18th and stopped new patient notes.
- Booking site certificate expiring on the 27th: renewed on the 14th;
would have shown a browser warning to every patient booking online.
- Backup of the records database failed on the night of the 20th:
target disk full; cleared and re-run on the 21st before 08:00;
would have left the clinic without a usable backup until noticed.
(source: work log, prevented-incident log)
2. WHAT WE HANDLED WITHOUT ANYONE NOTICING
Alerts acknowledged and resolved by a person: 23 (last month: 31)
Three worth a sentence:
- 3rd, 06:20: mail queue growing on mail.example.com; transport
service restarted; appointment reminder emails would have stopped
from 08:00.
- 11th: reception switch port 7 flapping for two days; patch lead
replaced; the check-in desk would have lost its terminal.
- 24th: UPS on the records server failed its battery test; battery
replaced on the 26th; the next power cut would have taken records
down without a clean shutdown.
(source: alert history)
3. WHAT WE CHANGED SAFELY
Maintenance windows: 4 of 4; follow-up alerts: 0
- Firewall firmware update: 2nd, 19:00-20:30; no alerts after.
- Records server patching: 16th, 20:00-22:00; no alerts after.
- Core switch second power supply: 16th; no alerts after.
- Booking site host patching: 23rd, 19:00-20:00; no alerts after.
(source: change record; alert history)
4. WHAT IS COMING
- Internet link: peak utilisation 68% and rising ~2%/month;
upgrade decision needed within four months.
- Records database volume: extension buys about ten months at current
growth; storage purchase to be in next budget.
(source: weekly trend review; capacity register)
AVAILABILITY THIS MONTH (planned downtime excluded)
Records: 100.00% Booking site: 99.97% Internet link: 100.00%
Method: monitoring checks every 60 seconds; planned windows excluded;
booking site was unreachable for 13 minutes on the 5th (host restart).
Next report: <first working day of next month>
Notice what the example does not contain. No mention of the hundred-odd alerts that cleared on their own. No list of tickets. No graph. Nothing described as critical or proactive. The reader learns that three outages were prevented, and that something went wrong most working days and was handled. They learn that four changes went in cleanly and that two decisions are coming. That is a month of work in under a page.
Producing it in under an hour
The report takes under an hour once the sources are in place, because every section maps to one place you already have. This is the order I use.
- Ten minutes: the work log. Read this month's lines. Every line with a specific consequence goes into "caught early". Every alert handled goes into the candidate list for section two.
- Fifteen minutes: the alert history. Count alerts acknowledged by a person this month. Then pull the list, sorted by how long each one was open, and pick the three whose consequence a reader will understand.
- Ten minutes: the change record. Count windows planned and completed. For each, look at the hour after it closed in the alert history and note whether anything fired.
- Ten minutes: the trend review notes. Take the two or three trends nearest a decision and write each as a line with a date.
- Five minutes: the availability line. Copy the numbers from the scheduled availability report. Do not recalculate them by hand; if the report is wrong, fix the report.
- Five minutes: read it aloud. Cut every adjective. Check every consequence names a service and a person. Send.
The alert history step is the one people find slowest, because they read the whole month. Do not. Ask the database for the acknowledged alerts sorted by time open, and read the top of the list. Table and column names differ between products, so treat this as illustrative and check the product documentation for yours:
-- Illustrative: adapt table and column names to your monitoring database
SELECT device, check_name, raised_at, acknowledged_at, cleared_at,
ROUND((julianday(cleared_at) - julianday(raised_at)) * 24, 1) AS hours_open
FROM alert_history
WHERE raised_at >= DATE('now', 'start of month', '-1 month')
AND raised_at < DATE('now', 'start of month')
AND acknowledged_at IS NOT NULL
ORDER BY hours_open DESC
LIMIT 20;
The header count is the same query with COUNT(*) in place of the column list. Keep both queries in a text file next to the template, and fifteen minutes becomes five.
Scheduled reports do most of the remaining work. If your monitoring product can email a monthly availability report and an alert summary on the first of the month, set that up once. Then the sources arrive before you have opened the template; scheduled delivery covers the mechanics. The quiet month report itself stays hand-written, because the three sentences and the consequences are judgement, and judgement is the part worth an hour.
Tip: write the "what is coming" section first. It is the section you are most likely to skip when tired, and the one your manager will remember, because it is the only one that asks something of them.
The report that changed the question
At Bellweather Logistics, the operations director asked the IT manager what the two administrators actually did. The question came in a budget meeting and not unkindly, since the depot systems had not gone down in living memory. The IT manager had no answer beyond "keep them up", and the headcount was frozen. The following month she sent the first quiet month report. It showed three catches, twenty-six alerts handled, five clean windows, and a warehouse scanner network that would run out of addresses by spring. The director's reply was four words long: "keep sending me these." The freeze was lifted two reports later, and nobody has asked what the administrators do since, because the answer arrives on the first of every month.
The report did not argue. It did not explain what monitoring was, or why prevention matters, or how hard the team worked. It listed what had been caught, handled, changed, and foreseen, in the shape it would use next month. And it let the reader draw the conclusion. "What does IT do?" is not answered by an answer. It is answered by a page that keeps arriving.
The checklist
Run through this before sending each month's report.
- All four section headings are present, even where a section says "none this month".
- Every "caught early" entry names the signal, the action with a date, and a specific consequence avoided.
- No "caught early" entry is an alert that cleared on its own.
- The handled count is alerts acknowledged by a person, not alerts raised, and last month's count sits beside it.
- The three sentences each name who outside IT would have noticed.
- Windows planned, completed, and followed by alerts are all stated as numbers.
- Every "coming" item has a date by which a decision is needed.
- The availability line comes from the availability report, excludes planned downtime, and states the method once.
- No adjectives, no raw alert count, no ticket list, at most one graph.
- The report fits on one page, and the next report date is at the bottom.
Where to go next
The quiet month is the hardest report to write because it has no story, and the fix is to stop looking for one. Four fixed sections, each sourced from the monitoring system, filled in plain words, in under an hour, on the first of every month. The reader learns the shape, then trusts the shape, and then stops asking what IT does, because the answer arrives before the question.
The sections draw on the rest of this pillar. Uptime reporting as reputation defence is the standing record behind the availability line. Answering "what do we pay you for?" is what twelve of these reports let you say in a single meeting. Both are part of Making Invisible Work Visible With Monitoring Data.
Put this into practice with SysPeep Monitor
The quiet month report is only quick if its sources are already in one place. SysPeep Monitor keeps alert history, availability reports, and scheduled reports in a single web console. A central Windows service and optional remote probes feed it. So the acknowledged-alert count and the availability line are a report away rather than a spreadsheet away. The free trial installs on a Windows machine if you want to write next month's report from your own data.
Download Free TrialLearn About SysPeep