Correlator

Function

Detects patterns and relationships between alerts from different services to identify systemic root causes.

Model

Claude Haiku

Availability

Only available on Team and Pro plans.

How it works

When a group receives a score > 30, the correlator analyses: - Active alerts on other services in the same namespace - Recent infrastructure events (nodes, network) - GitLab pipeline failures from the same repository

If a pattern is detected, the group is enriched with a correlations field:

{
  "correlations": [
    {
      "serviceId": "production/api-gateway",
      "reason": "HighLatency",
      "confidence": 0.87,
      "explanation": "api-gateway shows high latency since the same timestamp"
    }
  ]
}

Correlation example

A NodeNotReady on node-3 can automatically correlate with all pods running on that node, raising the score of each one and identifying the root cause as a node failure, not an application issue.