Solution · Incident management
Restore service faster, and stop the same incident coming back.
An incident is any unplanned interruption to a service your people rely on — email down, the customer portal timing out, the office network dropping. How fast you restore it, and whether you learn from it, decides whether the next one costs you an afternoon or a customer.
Every incident logged in one place, with a time and an owner.
Priorities set by urgency and impact, not by who shouts loudest.
Repeat incidents traced to a root cause and fixed for good.
The problem
Why incidents hurt more than they should
Most teams do not have an incident problem — they have a process problem. The technical fix is often the easy part; the delay comes from everything around it.
Nobody knows it is broken
Users notice the outage before the team does, and the first report arrives as a chat message, a phone call, or a walk to someone’s desk. Without one place to log incidents, the clock starts late and the record starts incomplete.
Everything feels urgent
Without a shared way to set priority, the loudest voice wins and the most important outage waits behind a password reset. The team works hard all day and still fixes the wrong things first.
The same fix, every week
A service fails, someone restarts it, and everyone moves on. Nobody records the cause, so the same incident returns next week, then the week after, taking the same hours to fix each time.
Silence while it is down
People can tolerate an outage; they cannot tolerate not knowing. When nobody sends an update, users open their own tickets and call the desk, doubling the work for a team already fixing the problem.
The process
Five steps, and the itserva page for each
Five steps, in order. Skip one and the next one costs more.
- 01
Detect and log
Record the incident as soon as it is known: what broke, who is affected, and when it started. One shared record beats five chat threads, and it is the record every later step depends on.
- 02
Prioritise
Score urgency against impact so a P1 outage jumps the queue and a minor glitch waits its turn. The same matrix, used the same way every time, ends arguments about what matters most.
- 03
Resolve within SLA
Work against the target your service level agreement sets, not against a feeling. Knowing how much downtime the SLA allows in a month tells you whether you have hours to fix this properly or minutes to restore service first.
- 04
Communicate
Tell the people affected what is happening and when the next update is coming. A short status note every half hour keeps one incident from becoming twenty duplicate tickets.
- 05
Review and prevent
For anything serious, ask why it happened and what would stop it happening again. The answer usually points at one fix — a missing monitor, a tired disk, a step nobody documented — and that fix is the real resolution.
What’s included
Tools, templates and guides for incident management
Template
Incident Report Template
Record what happened, who was affected, and what was done — in one page.
Product
Service Desk
Log every request and incident in one place, with priority and SLA timers built in.
Template
Root Cause Analysis Template
Work through the five whys and the contributing factors behind an incident.
Guide
Incident Management Process
From detection to closure: the steps every incident should go through, and the ones teams skip.
Measure it
The four numbers worth tracking
Four numbers tell you whether incident management is improving. Track them monthly, split by service, and investigate when one moves.
MTTR
Mean time to resolve: total downtime divided by the number of incidents. When it falls, your team is recovering faster; when it rises, ask whether the fix or the diagnosis is the slow part.
SLA met rate
The share of incidents resolved within the target your agreement promises. A falling rate is usually a staffing or prioritisation problem, not a technical one.
Volume by category
How many incidents each service and cause produces. A category that keeps growing is a problem wearing an incident’s clothes.
Repeat incidents
Incidents with a cause you have seen before. Every repeat is a fix that was never made, and the cheapest incidents to remove are the ones you have already solved once.
Frequently asked questions
What counts as an incident?
Any unplanned interruption to a service, or a reduction in its quality. A server down is an incident; a request for a new laptop is a service request, and mixing the two makes your incident numbers meaningless.
How fast should we respond to a P1?
As fast as your service level agreement promises — that is what the agreement is for. Agree the target with the people who rely on the service before the outage, not during it.
Do small teams need an incident process?
Yes, and a simple one. Log it in one place, prioritise it the same way every time, and review anything serious. That is the whole process; the rest can wait until the basics work.
When is an incident a problem?
When the same cause produces incidents more than once. Ten separate incidents with one cause are one problem, and treating them as ten incidents means the cause never gets fixed.
Start with one incident, logged properly.
Pick the next incident, log it in one place, and prioritise it with the matrix. The process builds from there — and a free workspace saves your work so the next review starts from facts.