If you follow how organisations keep their digital services running, 2026 has been a busy year for the corner of operations known as IncidentOps. Tools have been retired, regulators have tightened their reporting clocks, and a few long-standing habits around measuring incident response are being questioned. This is a friendly, plain-language tour of five shifts worth watching, written for administrators, managers and curious readers who work in organisations that depend on booking, records and messaging systems, hospital back offices included. It deals only with technology operations. Nothing here relates to clinical care, and nothing is advice for any health decision.
Our reference throughout is the Ops Atlas, a learning-focused index of operations disciplines. In the passage on IncidentOps as the Atlas describes it, the term is called an informal, rarely used label for incident management, with a snapshot of the tool landscape dated October 2026. Treat the shifts below as a reading of that snapshot, not as predictions.
Shift 1: The alerting tools are changing hands
The most concrete news is about software that wakes people up at night. The Atlas records that Opsgenie, a widely used on-call tool, shuts down on 5 April 2027, with data that has not been migrated deleted at that point; its owner is steering customers towards a service-management product instead. In a parallel move, the open-source edition of Grafana OnCall had its cloud connection switched off on 24 March 2026 and the project was archived, while Grafana’s hosted incident-response offering carries on. If your team uses either, the lesson is practical: check your migration plan now, export your schedules and escalation rules, and test the new setup with a drill well before any deadline.
For non-engineers, the key takeaway is simple. Any organisation whose appointment desk, records team or billing office depends on a paging tool should ask who owns that tool and whether anyone has written down how alerts reach a human.
Shift 2: Regulators are setting the clocks
The second shift is that outside rules increasingly define how fast an organisation must speak up. The Atlas points to the EU’s NIS2 framework, which expects a first warning inside a day of learning about a significant event, a fuller notification after about three days, and a closing report after a month. Whether any such rule binds a particular organisation is for lawyers to say, and this article does not try. The operational point stands, though: if a regulator’s clock starts when you become aware, then the habit of declaring an incident early is not only good manners, it creates a timestamp that everyone can work from.
The Atlas also notes that a US federal publication on security incident response, revised in April 2025, now sits in the background for attacker-driven events, which it keeps separate from the ordinary operational incidents covered here. In practice, many organisations run the two side by side, with the security workstream moving into a restricted space while the coordinator keeps the wider picture.
Shift 3: Averages are losing their shine
The third shift is about measurement. For years, managers asked for a single figure describing how quickly incidents were resolved. The Atlas shows why that number is slippery: the abbreviation behind it stands for several different things, and because incident times are lopsided, with a few very long events, one unlucky quarter can swing the average wildly. It cites a simulation in which the average moved by a fifth or more in about seven out of ten quarters despite no real change in performance.
The suggested replacement is less tidy but more honest: show the spread of times for each phase, track how much of the reliability allowance was used, watch how heavily paging lands on people, and count how many promised follow-up actions are actually finished. If your quarterly report still has one headline number for incident recovery, now is a good time to ask what it hides.
Shift 4: The human side is moving up the agenda
Fourth comes a rising interest in whoever is holding the pager. The Atlas cites a 2026 survey from a monitoring-tools company, with the caveat that its 1,363 respondents chose to take part, in which roughly three in ten pointed to alert fatigue and about one in six to painful coordination between teams. Self-selected samples are not representative, but the themes ring true. A related example in the Atlas shows that six engineers sharing a weekly rota with two round-the-clock slots would be on call for around a third of their time, above a one-quarter ceiling used by one large company, and that about eight people would be needed to stay under it.
If your organisation runs automated equipment as well as software, our overview of managing a robot fleet so performance stays steady explores a similar balance between machines and the people who watch them.
Shift 5: Roles over heroics
The last shift is cultural. The Atlas sketches a five-step maturity scale that begins with “heroic” effort, where results hinge on who happens to be around, and rises through a defined severity policy and rota, a kept communication rhythm, a learning stage in which most priority actions finish on time, and an adaptive stage. It frames these as its own proposal. What they share is a move from individuals to roles: the coordinator hat, the communicator hat, the hands-on hat and the note-taker hat, each worn by whoever is trained and available.
Small organisations do not need all of it. The Atlas suggests a ladder of tiers: begin with tickets, add one rota and a message template, and only add named roles and a public status page when incidents regularly cross teams or serious events happen twice a month. More elaborate structure with deputies and sub-teams matters only if more than five responders are regularly needed.
A recurring warning: do not let the messenger fail with the message
One theme runs underneath all five shifts, and it deserves its own mention. The systems that tell people about a problem must not depend on the thing that has the problem. The Atlas recalls a well-known 2017 case in which a large storage service stumbled and the provider’s own health dashboard, which relied on that same service, could not be used for hours. It also recalls the 2022 episode described above, where silence, not the technical fault alone, did the reputational damage. For any administrator, the quick check is easy to run: if our main system went dark tonight, how would we reach our users, and would that route still work?
A related caution from the Atlas concerns secrets. Configuration tools that mark a value as sensitive may only hide it from printed output while the value still sits in stored state, so routing keys for alerting services belong in encrypted storage. Asking the question takes no technical training, only curiosity, and the answer tells you a lot about how carefully the setup was done.
Looking across the five shifts, a pattern emerges. Tools come and go, rules get stricter and metrics get more nuanced, but the human basics stay put: decide early, name who is doing what, tell people on a rhythm, act on what you know, and learn without blame. An organisation that has those basics survives a change of vendor much more easily than one that has only a product. That is why the Atlas keeps its tool list short and its practice list long, and why a modest, rehearsed routine is worth more than an expensive dashboard nobody has practised with.
What to do about IncidentOps this quarter
- Inventory your alerting and status tools, and note any with an announced end date.
- Write a short list of the people who must be told, and how fast, when a system fails.
- Replace any single headline figure for incident time with a short table of ranges.
- Count the hours your on-call people actually carry.
- Run one rehearsal, even if it is a thirty-minute conversation round a table.
For readers who want to build the technical skills behind this, a structured practice plan for DevOps interview rounds offers a path into the surrounding engineering topics. The overall picture is of a field that is quietly maturing, even if the name itself remains casual. Revisit your own arrangements in six months, and see which shifts have reached your door.