Triage

Classify first, fix second.

Chapter 1 of 4

How Emergency Rooms Sort Chaos

The triage nurse’s role is to determine the urgency of the patient’s condition and decide how quickly they need to see a doctor. No one gets treated in order of arrival. They get treated in order of need.

This sounds obvious until you realize how many tech teams still process incidents the way a deli counter processes customers: first come, first served.


Priority Levels

Patients are classified into different levels of priority depending on their condition’s severity. The exact names vary by hospital and country, but the system generally follows five levels:

The key principle: urgency over arrival time. The most critical patients are treated first, regardless of when they walked in.


The Tech Translation

In tech, you can adopt the same classification for incidents and system issues:

Level Hospital Tech Equivalent
1 — Immediate Heart attack, respiratory failure Critical outage, security breach, data corruption
2 — Urgent Severe pain, large wounds Performance degradation, partial outage that could escalate
3 — Less Urgent Mild asthma, uncomplicated fracture Slow response times affecting some users, non-critical service degraded
4 — Non-Urgent Minor cuts, sprains Minor bugs not affecting core functionality
5 — Routine Cold, minor rash Cosmetic issues, low-impact bugs, tech debt items

Most teams already have something like this — SEV1, SEV2, SEV3 labels in their incident management tool. The problem is not the labels. The problem is that people treat them all the same way once the adrenaline kicks in.


Treatment Based on Priority

After classification, patients are treated based on their priority level rather than arrival order. A person with a life-threatening condition (Level 1) is seen immediately. Someone with a sprain (Level 4) waits.

In Level 1 and Level 2 situations, the team acts immediately — often multiple professionals at once. Levels 3 through 5 wait in a designated area or get redirected elsewhere.

In tech terms: a SEV1 gets the war room, the pager cascade, and cross-team coordination. A SEV4 gets a Jira ticket and a place in the backlog. When you treat every alert like a SEV1, you burn out your team and desensitize them to the real emergencies.


Continuous Reassessment

Triage is not a one-time process. Patients in the waiting area are continuously reassessed to ensure their condition is not worsening. If a patient’s condition deteriorates, they are moved to a higher priority level and treated sooner.

In tech terms: that “minor” memory leak you classified as SEV4 on Monday might be a SEV2 by Thursday if it is growing linearly. The status page API returning slow responses might be fine until a major incident hits and suddenly nobody can see the status updates.

Key principles:


A Triage Example

Imagine three incidents land at the same time:

The triage engineer ensures Incident A gets immediate attention and the full on-call response. Incident B is acknowledged, a workaround is noted (manual deploy if needed), and it is queued for after the crisis. Incident C gets a ticket.

If you have ever seen a team spend 20 minutes debating the deploy pipeline while production is burning, you have seen what happens without triage.


Chapter 1 of 4