Error Feed Triage Workflow: Resolve, Ignore, and Escalate
How to move issues through the Error Feed triage workflow: resolving, acknowledging, ignoring, assigning, and escalating.
About
Triage is reviewing new issues, deciding what to do with each one, and tracking them through to resolution. Error Feed is built to fit into your existing workflow rather than replace it: issues have statuses, assignees, and integrations that connect to whatever process you already use.
Status lifecycle
An issue moves through four primary states:
Unresolved → Acknowledged → Resolved
↕
EscalatingUnresolved is the starting state. Every new cluster starts here. Filter the feed to Unresolved to see what hasn’t been looked at.
Acknowledged means someone has reviewed the issue and confirmed it’s real and worth tracking. It doesn’t mean a fix is in progress; it means the issue has left the “inbox.” Use Acknowledged to reduce noise so your team knows what’s new vs. what’s known.
Escalating means the issue is actively getting worse or a previous fix didn’t hold. Use this to flag urgency beyond “this is open.” Issues can go straight from Unresolved to Escalating if they warrant immediate attention.
Resolved means the fix is deployed and the issue is closed. Error Feed keeps monitoring; if the same cluster pattern reappears, it creates a new issue rather than reopening the old one. This keeps resolved issues genuinely resolved and regressions visible.
Changing status
Three paths:
Header buttons: the fastest route. The issue detail header has Resolve, Acknowledge, and Ignore issue buttons. One click, no confirmation.
Status dropdown in the metadata panel: click the current status chip in the Triage section for a dropdown with all states. Use this for states like Escalating that don’t have a dedicated header button.
Feed list: status changes made in the detail view show up in the list view immediately. No reload needed.
Warning
“Ignore issue” currently sets the status to Escalating, which doesn’t permanently suppress the issue. To stop seeing an issue, set it to Resolved.
Assigning issues
Issues can be assigned to anyone in your organization. The assignee field is in the Triage section of the metadata panel.
Click Assign to open the picker, then pick a team member. To reassign, click the current assignee and pick a new one. To unassign, click the current assignee and select Unassign.
Assignment is informational right now: it doesn’t send a notification. For that, use the Linear integration to create a ticket and assign it there.
A practical triage session
Here’s how to work through a backlog of issues efficiently:
Filter to Unresolved + Critical
Start with the highest-severity issues nobody has looked at yet. Your “on fire right now” queue.
In the Feed list, set Status to Unresolved and Severity to Critical.
Open each issue and read the description
The Overview tab’s Description section says what’s going wrong in plain language. For most issues you’ll know within 30 seconds whether it’s a real problem or a false positive.
Acknowledge or resolve immediately if you can
Recognize the issue and already have a fix in flight: set it to Acknowledged and assign it to whoever owns the fix.
Already fixed from a previous deploy: set it to Resolved.
Not something you’re going to act on: still Acknowledge it so others know it’s been reviewed.
For complex issues, dig deeper
Use the Trends tab for severity and trajectory. Use the State Graph to see where in the workflow the failure concentrates. Use Deep Analysis when an issue warrants more investigation.
Create a Linear ticket for actionable issues
For issues that need a proper fix tracked in your project management tool, use the Linear integration to create a ticket directly from the metadata panel. The ticket is linked to the cluster, so you can navigate between them.
Repeat for High, then Medium severity
Work down the severity ladder. High after Critical, Medium after High. Low-severity issues can be batched into a weekly review instead of triaged one by one.
Handling escalating issues
An issue escalates when the problem is getting worse. In practice: watch the Trends tab Events Over Time chart. If the error count is rising week over week, move the issue to Escalating.
Treat Escalating issues like Critical regardless of their assigned severity. Severity is about the type of problem; Escalating is about the trajectory.
After a fix is deployed
When you ship a fix for a resolved issue:
- Set the issue to Resolved if it isn’t already.
- Come back 24–48 hours later and check the Trends tab. Confirm Score Trends are improving and the Events Over Time count is dropping.
- If a new issue appears in the same category with a similar name, the fix may have partially worked, or a regression was introduced. Investigate the new cluster.
The goal isn’t an empty issue list. It’s making sure nothing critical sits unreviewed and that resolved issues actually stay resolved.
Next Steps
Questions & Discussion