Fluidify

Alert Deduplication & Grouping

Regen prevents alert storms from creating hundreds of duplicate incidents. Two mechanisms work together: deduplication (same alert firing again) and grouping (different alerts that belong to the same incident).

Deduplication

When an alert arrives, Regen checks if an active alert with the same external_id already exists:

  • Match found — Regen updates the existing alert (timestamps, annotations). No new alert or incident is created.
  • No match — Regen creates a new alert and evaluates routing rules.

The external_id source depends on the integration:

Source ExternalID used
Prometheus fingerprint field
Grafana fingerprint field
CloudWatch AlarmArn
Generic external_id field you provide

Grouping Rules

Grouping combines multiple related alerts into a single incident. Configure under Settings → Grouping Rules.

Example: Group all alerts from the same service

Match labels:  service = payments-api
Group by:      service
Window:        5 minutes

If payments-api fires three different alerts within 5 minutes, Regen creates one incident and links all three alerts to it.

How it works

  1. A new alert arrives
  2. Regen checks active grouping rules in priority order
  3. If a rule matches, Regen looks for an open incident that already has alerts matching the same group key
  4. Match found — the alert is linked to the existing incident (no new incident created)
  5. No match — evaluated against routing rules to determine whether to create a new incident

Grouping rule fields

Field Description
Name Human-readable label
Match labels Conditions that must match for this rule to apply (key=value pairs)
Group by Label keys used to compute the group key
Window Time window — alerts in the same group that arrive within this window are linked to the same incident
Priority Lower number = evaluated first

Routing Rules

Routing rules control what happens after deduplication and grouping. Configure under Settings → Routing Rules.

Actions

Action Description
Create incident Automatically create an incident for matching alerts
Notify only Record the alert but do not create an incident
Suppress Silently drop the alert (useful for maintenance windows)

Example rules

Create incidents for critical and warning alerts:

Match: severity IN [critical, warning]
Action: Create incident

Suppress info alerts during business hours:

Match: severity = info
Action: Suppress

Rules are evaluated in priority order. The first matching rule wins.