Support work has a particular kind of tiredness. It is not the tiredness of a long project; it is the tiredness of never finishing a thought. A ticket arrives, then a chat message, then an automated warning, then a colleague leans over with a quick question. By the end of the shift you have touched everything and completed little. This how-to follows a fictional support team through the steps of setting notification quiet hours, so that you can copy the process and adapt it to your own desk.
Meet the team
The team in our story looks after hosting and deployment for a set of client applications. There are five of them, working staggered shifts, sharing a ticket queue, a chat workspace and an alerting tool. They are good at their jobs and exhausted by their days. Their lead, who nobody ever hears raise a voice, decides it is time to change the way notifications work, and starts by asking everyone to describe their worst hour of the previous week.
Step 1: Describe the worst hour
The answers come quickly. One engineer was halfway through diagnosing a slow database when four chat mentions arrived, each reasonable, and by the time they replied to all of them the original trace had gone cold. Another was writing a postmortem for a past incident and was interrupted so often it took two days. A third simply said: “I never get through the queue in order.”
Starting here matters. People are more willing to change a habit when they have just described the pain it causes in their own words, and the stories give the lead a vocabulary to use later when explaining the plan to clients and managers.
Step 2: Sort what arrives
Next the team lists every source of interruption and sorts each into three groups. The first group is things that mean a customer is affected right now. The second is things that need an answer the same day. The third is everything else.
The sorting reveals something useful. The first group is tiny. Most of the daily noise belongs in the second or third group, which means most of it can wait for a few hours without anybody coming to harm. That insight is the foundation of the whole arrangement.
Step 3: Choose the windows
The team looks at their shifts and finds the times when the queue is calmest and the hardest tasks, such as root-cause analysis and runbook writing, are usually attempted. They choose two protected windows per shift, each roughly ninety minutes, separated by an open period for catching up. Because shifts are staggered, the windows are staggered too, so at any moment at least two people are fully reachable.
This detail deserves a note for your own planning. A support team cannot all go quiet at once. Stagger the windows so coverage is continuous, and make sure the person who is open during somebody else’s window is clearly identified.
Step 4: Write the exception route
This is the step the lead insists on doing before anyone mutes a thing. The rule is written on one page. It says that a confirmed outage affecting customers, a security alert and a direct call from the escalation contact are all reasons to break a window. Everything else waits.
The rule also names the route. Chat is not the route, and neither is email. A phone call to the person on point is, because calls are hard to send idly and hard to miss. The rule lists the on-point person for every window of every shift, in a table pinned to the top of the team channel.
Step 5: Route alerts by lane
With the human rules written, the team turns to the tooling. In their alerting tool, they define three lanes. A page rings the on-point engineer. A chat post lands in a channel that is read at set times. A ticket is created for slow-burning issues.
They go through existing alerts one by one. A service returning errors to customers becomes a page. A disk filling slowly over several days becomes a ticket. A certificate expiring next month becomes a ticket too, with a due date. Several alerts that had been paging people at night turn out to be informational, and the team moves them down. Only paging alerts will ever cross the boundary of a quiet window.
Step 6: Set the devices
The technical settings are the shortest step. On phones and laptops, each person schedules do-not-disturb to match their windows, allows calls from the on-point contact and the lead, and keeps the pager app’s critical alerts switched on. Chat badges are muted, email previews are off and the lock screen shows no message text.
A small point of etiquette helps. Everyone changes their chat status when a window starts, with a short line such as “heads-down until the break; call for outages”. The status costs nothing and answers the question before anyone asks it.
Step 7: Plan the catch-up
Muted notifications are only harmless if somebody reads them. The team schedules a ten-minute catch-up at the end of each window. During it, each person clears the chat backlog, answers what can be answered and promotes anything that turns out to be urgent. The ticket queue is reviewed in the open period.
The lead adds a final rule: if you notice something non-urgent during a window, jot it on a shared list rather than sending a message. The list is read in the open period, and the habit stops people from adding to each other’s interruptions.
Step 8: Try it for two weeks and review
The team starts on a Monday and counts the calls it gets. How often did the exception route ring? How many were genuinely urgent? Did any urgent item arrive late?
At the review, the tally shows that most calls were justified, two were avoidable and one genuine problem had waited too long because it had arrived through chat. The team adds that source to the page lane and reminds everyone that chat is never the route. Nobody argues, because the tally has done the arguing.
What changed
The change is gradual and mostly unremarkable. Tickets close in a calmer rhythm. The postmortem that used to take two days takes an afternoon. Conversation in the open periods becomes warmer, because people are no longer irritated by constant interruption. The lead keeps the rule on a single page and revisits it every quarter.
If you are putting together your own support practice, background reading may help. An article on building a strong DevOps engineer portfolio covers the sort of documentation that quiet time allows, and a piece on the different role titles and career paths in DevOps explains the people who typically sit on a support rota.
A word on handling clients
Support teams answer to customers, so the lead also writes a short note for them. It explains that engineers work in focused blocks, that tickets are triaged on a regular rhythm and that confirmed outages are handled immediately by a named person reachable by phone. The note contains no apology and no jargon; it simply sets out how to reach the team and what to expect.
Clients rarely object. Several say it is the first time they have understood what happens between sending a ticket and receiving a reply. One even asks whether their own staff could adopt the same pattern, which is a flattering sign that the arrangement is understandable from the outside.
Keeping the rule alive
Rules fade when nobody tends them. The lead puts a recurring item on the team’s quarterly agenda: read the one-page rule aloud, ask whether the windows still match the work, check the on-point table for stale names and look at the tally of exception calls. The meeting takes fifteen minutes, and it prevents the slow drift that otherwise turns a clear arrangement into a half-remembered custom.
One last tangent
Some of the team study in their free time, and not always on technical topics. One of them browses Wingversity, a site laying out a web learning platform on the theme of aviation careers, which calls itself a preview filled with sample content. They apply the same rule at home: phone muted for the study hour, with calls allowed from family only.
Your checklist
- Ask each person to describe their worst interrupted hour.
- Sort incoming signals into urgent now, same day and later.
- Choose staggered windows so coverage stays continuous.
- Write the exception route on one page and name who is on point.
- Define page, chat and ticket lanes in your alerting tool.
- Schedule do-not-disturb and set status messages.
- Plan a short catch-up after each window.
- Trial for two weeks, tally the calls and adjust.
The steps are ordinary, and that is the point. Support teams do not need a transformation; they need a shared agreement about when to be reachable and a trustworthy route for the exceptions. Write it down, try it, and let the tally tell you what to change.