🔔 Incident Grouping V2 — Now in Private Preview

What's new

We've significantly improved the logic that groups related monitor failures into the same incident. Two changes ship together:
Smarter matching rules across monitor types. Previously, grouping was too narrow: many logically related failures (e.g. a Volume failure and a Freshness failure on the same asset) weren't grouped together. We've rebuilt the matching matrix from the ground up, now covering all monitor templates with rules tuned to the direction of the relationship:

• Upstream matches: failures that could have caused the incident
• Downstream matches: failures likely caused by the incident
• Same asset matches: co-occurring symptoms on the same dataset

Sibling asset support. Grouping now extends to sibling assets: tables that share a common upstream parent, or that are both root tables in the same datasource/schema. Sibling failures are strong co-symptoms of a shared root cause. Priority order is now: Same asset > Upstream > Siblings > Downstream.

Why it matters

Related failures that previously created separate incidents will now be consolidated, reducing noise and making it easier to investigate the true root cause.

Status: Private preview — gated by feature flag. Reach out to enable it for your instance.