Quick summary

What Linear Triage is: a per-team inbox in front of a Linear team's workflow. Issues from integrations, Linear Asks, and people outside the team wait there until a person reviews them.

How you clear it: accept (1), decline (2), mark as duplicate (3 or MM), or snooze (H). Only accept moves an issue into the backlog.

How to automate it: on Business and Enterprise, Triage rules apply set actions as issues arrive, Triage Intelligence suggests an assignee, labels, and duplicates with AI, and Triage responsibility notifies or assigns an owner.

Connect a few integrations to a Linear team and the backlog stops being a plan. Every Slack escalation, every Sentry alert, and every issue a teammate from another team files lands in the same list as the work you actually committed to this cycle. The signal drowns. Linear Triage is the fix: a holding inbox that catches everything incoming and keeps it out of the real backlog until a person decides it belongs there.

This is a practical guide to the feature, not the general practice of triaging bugs, which is covered in the bug triage guide. Here we cover what Triage is, what routes into it, the four ways to clear an issue out of it, the three features that automate it (Triage rules, Triage Intelligence, and Triage responsibility), and how to feed it cleanly from customer support.

In this article

1.

2.

3.

4.

5.

6.

7.

8.

9.

What Linear Triage Is

Triage is a special inbox that sits in front of a team's workflow. Issues that arrive from outside the team collect there instead of dropping straight into the backlog, and each one waits for a person to review it and decide whether it becomes real work. It is enabled per team, so a support-facing team can run it while a team that only generates its own work does not have to.

Triage moves one decision out of the backlog: does this belong in our work at all. Made by accident, that decision fills the backlog with noise. Made on purpose, it keeps the backlog something the team can trust.

The reason this matters is that a backlog only works as a plan if everything in it was put there deliberately. The moment integrations and other teams can write directly into it, it becomes a record of everything that happened rather than a list of what the team chose to do. Triage restores the boundary.

What Lands in Triage, and What Does Not

Triage catches work that originates outside the team. Three sources route into it.

  • Issues from integrations

    Slack, Sentry, and the other connected tools file into Triage rather than straight into the backlog, so an automated alert or a forwarded escalation does not jump the queue ahead of planned work.

  • Issues from outside the team

    When a workspace member who is not on the team creates an issue for it, that issue lands in Triage. The team reviews it before it counts as their work, instead of finding it already sitting in the cycle.

  • Issues created in the Triage view

    Anything you add directly while working the queue starts in Triage, which keeps it consistent with everything else waiting to be reviewed.

Linear Asks is the main intake route for people who do not use Linear. It turns requests from Slack, email, and web forms into Linear issues, and Slack Asks created for a team start in that team's Triage. Asks is a Business and Enterprise feature; the Linear Asks guide covers each channel.

What does not land in Triage is the team's own normal work. Issues that team members create through the usual flow go straight to the team's states, and a default template in Team Settings can also set a status that overrides Triage. Triage is specifically for work arriving from elsewhere, which is exactly the work most likely to be unvetted.

How to Turn Triage On

Triage is off until you enable it, and you enable it per team. Open Team Settings, go to the Triage section, and toggle it on. A Triage entry appears under the team name in the sidebar, and you can jump to it from anywhere with the G then T shortcut. From another team's views, O then T lets you pick the team first.

Two settings are worth knowing on day one. You can require a priority to be set before an issue leaves Triage, in the same Triage settings. And Triage issues are left out of views by default, so if you want them in a custom view, add Triage to its status filter.

The Four Ways to Clear a Triage Issue

A triaged issue is not done until it leaves the queue, and there are exactly four ways out: accept, decline, mark as duplicate, and snooze. Each has a keyboard shortcut, because Triage is meant to be worked quickly. If you need more information first, comment on the issue and leave it in Triage, or snooze it.

  1. 1

    Accept (shortcut 1)

    Accept moves the issue to the team's default status, which is the only action that promotes it into the real workflow. Set the assignee, priority, and project as you accept so it lands ready to work rather than as a bare title that has to be triaged a second time later.

  2. 2

    Decline (shortcut 2)

    Decline moves the issue to a Canceled status, with an optional comment explaining why. A lot of declines are not defects at all, they are feature requests that need a different home and a reply to the person who asked. The feature request system guide covers where those should go instead.

  3. 3

    Mark as duplicate (shortcut 3 or MM)

    Point the issue at the one it repeats. Its attachments and customer requests move onto that issue, and the duplicate is canceled. This is where the many-reports, one-issue pattern pays off: ten customers reporting the same bug collapse onto a single tracked issue instead of ten, so engineering hears the problem once and sees how many accounts it affects.

  4. 4

    Snooze (shortcut H)

    Snooze hides the issue until a time you pick or until new activity arrives on it, whichever comes first. It is for work that is real but not now: you do not want it cluttering the queue, and you do not want to cancel it. Snoozed issues are hidden from teammates by default too, and a View Options toggle shows them.

Automating Triage: Rules, Intelligence, and Responsibility

Working the queue by hand is fine at ten issues a week. Past that, Linear has three features that do part of the work. All three are on the Business and Enterprise plans, so check your plan in the Linear pricing guide before you design around them.

FeatureWhat it doesWhere you set it up
Triage rulesApply set actions to matching issues as they enter TriageThe team's Triage settings
Triage IntelligenceAI suggestions for team, project, assignee, labels, and duplicatesWorkspace Settings, then AI (admin)
Triage responsibilityNotify or auto-assign the person on triage dutyThe team's Triage settings

Triage Rules

Triage rules take predefined actions on issues when they enter Triage. Each rule is triggered by filterable properties, and it can set the issue's team, status, assignee, label, project, or priority. A rule might send anything labeled Billing to the payments team, or put every issue from a key customer at Urgent.

Rules run in order from top to bottom. If a rule moves an issue to another team's Triage, that team's rules apply too, and Linear shows conflicts between rules in the settings. To match any of several values, for example any of three customers, hold Shift while picking each value in the filter menu.

Triage Intelligence

Triage Intelligence is Linear's AI for the queue. Linear describes it as agentic AI that analyzes every new issue in Triage against your existing issues and your workspace's history. It suggests properties (team, project, assignee, and labels) and surfaces likely duplicates and related issues when an incoming issue is strongly similar to one you already have.

  • Turning it on

    An admin enables it under Settings, then AI. That turns it on for every team. Each team can switch it off, or limit suggestions to its own team and sub-teams.

  • Show, hide, or auto-apply

    Per team and per property type, you choose whether suggestions appear, stay hidden, or are applied automatically, optionally only for specific values. Sub-teams inherit the parent's settings and can override them.

  • Speed and guidance

    Suggestions take about one to four minutes to appear on a new Triage issue. You can add written guidance at the workspace, team, or sub-team level, and the most local guidance weighs most. For issues outside Triage, run Find Suggestions from the command menu.

Plans below Business still get lighter quick suggestions in the issue composer and property menus, which Linear calls much less thorough. Triage Intelligence is a different thing from Linear Agent, the assistant you chat with; the Linear Agent guide explains the difference.

Triage Responsibility: Who Owns the Queue

A queue with no owner becomes a graveyard. Triage responsibility defines who handles incoming issues, so the inbox always has someone accountable for clearing it.

  • Notify or auto-assign

    In the team's Triage settings, pick the members who handle incoming issues, then choose whether new issues notify them or are assigned to them automatically. Teammates creating an issue can see who is on triage duty.

  • Rotate it from your on-call schedule

    Triage is interrupt-heavy, so leaving it on one person forever burns them out. You can connect a PagerDuty, OpsGenie, Rootly, or Incident.io schedule to rotate the first responder. Other providers can build a custom schedule with Linear's API.

  • Pair it with your support rotation

    If you run an engineer-on-support rotation, the engineer on duty is the natural triage owner, since they already hold the week's customer context. The engineer-on-support rotation guide covers that setup in depth.

Triage Versus the Backlog

Triage holds incoming, unreviewed issues, and the backlog holds issues the team has accepted. The two get confused because both hold issues that are not being worked yet. The difference is commitment, and it changes who the list is for.

Triage (incoming, unreviewed)

  • Holds issues from integrations and outside the team
  • Nothing here is committed work yet
  • Excluded from normal views and reports by default
  • Each issue waits for accept, decline, duplicate, or snooze
  • Answers: should this become work at all?

Backlog (accepted, planned)

  • Holds issues the team has accepted
  • Committed to the workflow, just not started
  • Shows up in normal views, cycles, and planning
  • Has an owner and priority once you set them
  • Answers: when do we do this?

The practical rule is that nothing should reach the backlog without passing through a person first. Triage is that person's queue. Accepting an issue is the moment it crosses from someone else's input into the team's plan, and keeping that crossing deliberate is the entire reason the feature exists.

Feeding Triage From Customer Support

Triage is built for work that arrives from outside the team, and nothing arrives from further outside than a customer. A bug reported in a support ticket is the textbook case: it is unverified, it needs a human accept or decline call, and it should not jump straight into a cycle. The catch is context. A triager clears an issue in seconds when the report carries reproduction steps, the affected account, and a link back to the conversation. They stall when the issue is a one-line title and they have to go chase the reporter before they can even judge it.

Linear's own support integrations cover Intercom, Front, Zendesk, and Slack, so those teams can file or link issues from the support tool. HubSpot is not on that list. If support runs on HubSpot Service Hub and engineering runs on Linear, this is where IssueLinker fits. From the HubSpot ticket, support turns a customer report into a Linear issue in one click, with the ticket context carried over, so it arrives in the engineering team ready to triage instead of as a bare title. Comments and notes sync both ways, so a triager's question reaches the support rep on the ticket without anyone leaving their tool, and when the issue is accepted, declined, or shipped, the new status is written onto the ticket so support can close the loop with the customer. The full setup is in the Linear HubSpot integration guide.

Send support issues to Triage with the context already attached

If support runs on HubSpot Service Hub and engineering runs on Linear, IssueLinker turns a ticket into a synced Linear issue in one click, with the ticket context attached, two-way comment sync, and the Linear status written back onto the ticket so support can tell the customer what was decided.

What to Do This Week

If your backlog has quietly filled with work no one chose, three steps start fixing it without changing how the team plans.

  • Turn Triage on for your highest-inflow team

    Start with the team that fields the most integration and cross-team issues. That is where the backlog noise is worst and the payoff from a review step is biggest.

  • Give the queue an owner and one rule

    Name an owner this week and put the duty on a rotation. On Business or Enterprise, set up triage responsibility to notify them, add one triage rule for your most common routing decision, and turn on Triage Intelligence with suggestions shown, not auto-applied.

  • Route support intake through Triage, not the backlog

    Make customer-reported issues land in Triage with their context attached, so engineering accepts them on purpose instead of discovering them already mixed into the cycle.

Triage does not slow a team down. It takes one decision, does this belong in our work, out of the backlog where it was being made by accident, and puts it in a queue where it is made on purpose. That is the whole feature, and on any team with real inbound, it is the difference between a backlog you trust and a list you have given up reading.

Frequently Asked Questions