The error-outlier model is joining failures that share a symptom but not a cause.
Fast triage depends on broad grouping, while useful investigations require clusters that remain causally coherent.
Who you’d be doing this for
“I opened twelve traces from the same cluster and half were a totally different timeout path.”
Charles Henderson · Software Engineer
Owns an API service and uses automated investigation clusters during incident response.
What is at stake
Error-cluster precision fell from 91% to 76% after a traffic-pattern change. You have to restore useful grouping while protecting recall for real incident patterns.
Why it isn’t already fixed
Every obvious fix costs something else. That’s the part you’d have to decide.
- cluster precision vs. error coverage
- fast mitigation vs. root-cause correction
- representation richness vs. inference latency
Why Datadog
Error outliers in Datadog APM guide engineers from an incident symptom into traces and logs.
Written with these in mind
Not your kind of problem? 40 more at Datadog, or browse every organization.
This is the setup. The work is inside.
Running it puts you in the room: the full situation and its constraints, stakeholders who push back in their own words, and the decisions that are yours to make. What you produce becomes a Day One Plan — work you can show someone instead of describing.