$100 Website Offer
Get your personal website + domain for just $100.
Limited Time Offer!
Claim Your Website NowA Legal News Desk and the Quiet Craft of ITOps
The morning summary of court and policy developments goes out at the same time every day, and readers treat that regularity as part of the product. Behind it sits a small team that rarely appears in the byline: the people who keep the publishing desk’s computers, accounts, shared drives and content system running. In IT circles their work is called ITOps, and a fictional newsroom story is a useful way to see what it actually involves. What follows is an illustration of how such a desk might work, written for general information only; it is not legal advice and describes no real organisation.
Monday: the desk that never explained itself
Picture a six-person editorial team at a legal-information publisher. Two editors, three writers and one coordinator produce a daily roundup of developments from courts and ministries. They also handle reader emails, a subscriber list and a content management system that a part-time contractor once configured. Nobody on the team would call themselves “IT”. Yet when the login page stopped accepting passwords at 7:10 one Monday, the coordinator, Meera (a made-up name for this story), became the de facto operations function.
She did what most small teams do: messaged the contractor, tried a reset, and told the writers to email drafts around in the meantime. By 9:00 three versions of the same roundup existed in three inboxes. The login was restored by noon, but the real damage was the version confusion, and nobody could say who had decided which draft was current. That muddle is the first thing ITOps exists to prevent.
The Ops Atlas from xOps School frames the function plainly. According to the Atlas entry on ITOps, it is the group that keeps IT services running by handling requests and incidents, removing causes of repeat failures, controlling change and keeping a map of what the estate contains. Four verbs, and the Monday scenario above touched every one of them.
Requests and incidents are different animals
A forgotten password is a request: a routine, predictable ask with a known answer. A login page that rejects everyone is an incident: something that was working and now is not. Lumping them together meant the contractor received a vague message and had to guess how urgent it was.
A tiny desk can borrow the sorting habit without buying any software. Keep two lanes on a shared board. Requests carry a name, a date wanted and who approved them. Incidents carry a start time, the symptom in the reporter’s own words and a single owner. Within a week the team in our story could see that most of what they called “emergencies” were requests that had waited too long.
Priority was the next puzzle. A sample formula quoted in the Atlas for one popular platform sets priority as impact plus urgency minus one. You do not need the platform to use the idea. Impact asks how many readers or colleagues are affected; urgency asks how soon the harm grows. The login outage hit all six people and every minute delayed the daily publication, so it ranked at the top. A single writer’s broken keyboard shortcut did not.
Why the same fault came back in August
Two months later the login failed again, and this time Meera had a spreadsheet. It showed that both outages followed a certificate renewal that nobody on the team had known was due. This is the dull, valuable heart of problem management: stop treating the second incident as a fresh surprise, and ask what the two have in common.
The fix was undramatic. A renewal date went on the team calendar, the contractor got written instructions, and a short note recorded what had broken and why. That note became the first entry in a small knowledge folder. The Atlas places knowledge that lives inside the ticketing tool within ITOps scope, though it assigns the wider method of writing and curating such articles to a different discipline, so the folder stayed deliberately modest.
Readers who follow this publisher’s daily legal roundups, such as the edition covering developments on 17 June, never see any of this. That is the point. Good operations is mostly invisible until it fails.
Change without drama
Then the publisher decided to move to a new template for the roundup page. The contractor proposed switching it over on a Friday afternoon, which would have been fine except that a tax-season special was scheduled for the next morning. This is where change control earns its keep, and where the newsroom nearly learned the lesson the hard way.
The Atlas lists three kinds of change. A standard change is pre-authorised and low risk, like adding a new writer’s account by the usual recipe. A normal change is reviewed first. An emergency change cannot wait, for instance a fix for a weakness attackers are already exploiting. The template switch was a normal change, and assessing it exposed the clash with the Saturday special in about five minutes.
A worked example in the Atlas shows how a written risk score can route a change. In that lab, a change coded CHG-302 scored 6 points, made of 3 for the tier of the system, 2 for a database schema alteration and 1 for falling outside the agreed window, and that total sent it to the change advisory board. The newsroom’s version was far smaller: a three-line checklist with a number beside each line. Anything scoring above a threshold the team chose went to a ten-minute huddle; everything else just went ahead.
Notice what a sensible process avoids. The Atlas lists, among typical anti-patterns, treating every change the same and adding process only as a reaction to a failure. The Friday near-miss tempted the team to demand sign-off from all six people for every tweak. Meera resisted. A formal approval round for adding a writer’s account would have slowed the desk to a crawl and taught everyone to bypass it.
The clock that kept running over the weekend
Because the publisher had started selling its roundup to a handful of institutional subscribers, it also needed promises: how quickly would a problem be answered? Here the newsroom ran into service levels and discovered something the Atlas highlights as a common trap, namely clocks whose calendar and pause rules nobody understands.
The Atlas offers a lab where one incident, INC-202, ran for 92.5 hours of wall-clock time yet only 4.5 hours counted against its target, because weekends and a holiday paused the clock. The newsroom took a simple lesson from this: write down when the clock starts, when it stops and which days it ignores, before promising anything to a subscriber.
A customer-facing promise is stitched together from internal ones, and the Atlas shows the sum. Our fictional publisher had only a contractor and a hosting company, so its chain had two links. Adding the numbers revealed that the promised response time was shorter than the hosting company’s own commitment, the kind of quiet contradiction that arithmetic catches.
Knowing what you actually own
In the autumn the team tried to answer a simple question: which devices, accounts and subscriptions does the publication depend on? It took two days. A laptop nobody remembered lending out, a domain registered under a former employee’s email, a newsletter tool paid for on a personal card. The Atlas treats a configuration database as part of the discipline, and its health lab, covering ten configuration items, found nine of them complete, two stale for more than sixty days, one orphan and one duplicate serial number.
Meera’s list was far from a full database, and she made a point of stopping early. The Atlas is explicit that such a register should hold what decisions require rather than everything imaginable. Her six columns were: item, owner, renewal date, where the login lives, what breaks if it vanishes, and when the row was last confirmed. A quarterly check that looked for stale rows took less than an hour.
Words that get blurred when shopping for tools
Shopping for software exposed a vocabulary tangle. The Atlas separates ITOps (the function and its people) from ITSM (the practices) and ITOM (the technical tooling that watches systems), and notes that vendors disagree on the boundaries. For a six-person desk, the sensible response was to stop debating terms and write down who does what.
What the story leaves for your own desk
If you run any small team that depends on technology, the newsroom’s year suggests a short list of habits. None of it needs a purchase order.
- Separate requests from incidents on a shared board, with a named owner for each.
- When a fault returns, look for what links the occurrences before fixing it a second time.
- Score changes on a tiny checklist and reserve meetings for the high scorers only.
- Write the start, stop and pause rules for any response-time promise.
- Keep a short register of what you depend on, and trim rows that nobody can confirm.
Elsewhere on the site, a reader who wants a sense of how the publisher frames its daily coverage can browse the roundup dated 19 June, which shows the kind of output this operational groundwork protects.
The fictional Monday ended better than it began. The roundup went out forty minutes late, and the following Monday it went out on time with nobody noticing, which is the most reliable sign that operations are working. Anyone considering these habits for their own team should treat this story as an illustration rather than a prescription, and adapt it to the tools and obligations they actually have.